JJG IP Holdings, LLC

Patent Trials and Appeals Board

Oct 4, 2021

IPR2020-00899 (P.T.A.B. Oct. 4, 2021)

Trials@uspto.gov Paper 45
571.272.7822 Date: October 4, 2021

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
____________

MACHINEX INDUSTRIES, INC.,
Petitioner,

v.

JJG IP HOLDINGS, LLC,
Patent Owner.
_______________

IPR2020-00899
Patent 9,950,346 B2
_______________

Before GEORGE R. HOSKINS, MICHAEL L. WOODS, and
ROBERT L. KINDER, Administrative Patent Judges.

WOODS, Administrative Patent Judge.

JUDGMENT
Final Written Decision
Determining All Challenged Claims Unpatentable
35 U.S.C. § 318(a)
ORDER
Denying Petitioner’s Motion to Exclude
Dismissing Patent Owner’s Motion to Exclude
37 C.F.R. § 42.64

IPR2020-00899
Patent 9,950,346 B2

2

I. INTRODUCTION
Petitioner, Machinex Industries, Inc., filed a Petition (Paper 1, “Pet.”)
requesting inter partes review of claims 1–16 (“the challenged claims”) of
U.S. Patent No. 9,950,346 B2 (Ex. 1001, “the ’346 patent”).1 Pet. 4. We
issued a decision to institute an inter partes review of these claims. Paper 11
(“Institution Decision” or “Inst. Dec.”).
After institution, Patent Owner, JJG IP Holdings, LLC, filed a Patent
Owner Response (Paper 19, “PO Resp.” or “Response”), to which Petitioner
replied (Paper 28, “Pet. Reply” or “Reply”). Patent Owner also filed a Sur-
Reply to Petitioner’s Reply. Paper 29 (“PO Sur-Reply” or “Sur-Reply”).
Petitioner further filed a Motion to Exclude several of Patent Owner’s
exhibits (Paper 34), to which Patent Owner opposed (Paper 35), and
Petitioner Replied (Paper 37).
Patent Owner also filed a Motion to Exclude one of Petitioner’s
Exhibits (Paper 43), to which Petitioner opposed (Paper 44).
Oral argument, or hearing, was held on July 21, 2021, and the
Transcript of the hearing has been entered as Paper 42 (“Transcript” or
“Tr.”).
We have jurisdiction under 35 U.S.C. § 6. Petitioner bears the burden
of proving unpatentability of the challenged claims, and the burden of
persuasion never shifts to Patent Owner. Dynamic Drinkware, LLC v. Nat’l
Graphics, Inc., 800 F.3d 1375, 1378 (Fed. Cir. 2015). To prevail, Petitioner
must prove unpatentability by a preponderance of the evidence. See 35

1 Petitioner identifies itself and Machinex Technologies Inc. as the real
parties-in-interest. Pet. 3. Patent Owner identifies itself as the sole real
party-in-interest. Paper 4, 1.
IPR2020-00899
Patent 9,950,346 B2

3

U.S.C. § 316(e); 37 C.F.R. § 42.1(d). This Final Written Decision is issued
pursuant to 35 U.S.C. § 318(a) and 37 C.F.R. § 42.73. For the reasons that
follow, Petitioner has shown that claims 1–16 of the ’346 patent are
unpatentable.
A. Related Proceedings
The parties identify JJG IP Holdings, LLP v. Machinex Indus., Inc.,
No. 1:19-cv-00437 (M.D.N.C.) as a related proceeding. Pet. 3; Paper 4, 1.

B. The ’346 Patent (Ex. 1001)
The ’346 patent is titled “Method and Apparatus for Sorting Recycled
Material.” Ex. 1001, code (54). The patent describes an apparatus and
method for sorting heterogeneous material, such as a mix of recycling
material. See id. at code (57), 1:14–24. To illustrate an embodiment of the
described system, we reproduce its Figure 1, below:
IPR2020-00899
Patent 9,950,346 B2

4

According to the ’346 patent, Figure 1 “is a front perspective drawing
illustrating a system for the separation of heterogeneous material.” Id. at
3:36–38. In particular, conveyer 1 carries material to identification camera
enclosure 4, which encloses a camera (not shown). See id. at 4:1–4.
“Disposed proximately” to the camera is array of lights 5, which highlight
the target area for the camera. Id. at 4:4–5. In one embodiment, the camera
is a hyperspectral camera, in which cameras and sensors “collect and then
combine spatial and spectral information” as a set of images. Id. at 4:6–9.
Computer system 3 is programmed to receive input from the camera and to
provide output signals to ejector nozzles, which “eject desired product.” Id.
at 5:5–9. The force from the ejector jets “propels the picked items into”
IPR2020-00899
Patent 9,950,346 B2

5

receiving hood 7, which is located above conveyer belt 1, where the items
are pushed toward collection point 9. Id. at 5:23–28.

C. Illustrative Claims
Of the challenged claims, claims 1 and 14 are independent. Ex. 1001,
5:66–8:17. We reproduce claims 1 and 9 below, reformatted from the
version provided in the ’346 patent to include bracketed alphanumeric
nomenclature that corresponds with Petitioner’s modifications or
nomenclature. See Pet. 40–51.
1. [Limitation 1(p)] A system for the identification and
sorting of heterogeneous material, the system comprising:
[Limitation 1(a)] a hyperspectral identification system for
capturing spectra of material,
[Limitation 1(b)] said hyperspectral identification system
comprises at least one hyperspectral camera said hyperspectral
camera configured to receive spectral data from a plurality of
selected spectral bands of infrared and visible light
corresponding to spectral signatures of target materials to be
identified if present in said heterogeneous material and spacial
data locating a position of said heterogeneous materials on a solid
belt;
[Limitation 1(c)] said camera disposed proximate to a
visible or infrared light source,
[Limitation 1(d)] said light source and said at least one
hyperspectral camera are disposed on the same side of a stream
of said heterogeneous material;
[Limitation 1(e)] said solid belt having first and second
ends, said belt running beneath said hyperspectral identification
system and upon which said heterogeneous material is conveyed
from said first end to said second end,
IPR2020-00899
Patent 9,950,346 B2

6

[Limitation 1(f)] said belt traveling at a preset rate, said
hyperspectral identification system being disposed over said
second end of said belt;
[Limitation 1(g)] a computer configured to receive and
analyze data from said hyperspectral identification system,
[Limitation 1(h)] to identify target materials of a first user
defined category from among said heterogeneous materials,
[Limitation 1(i)] identify the spatial position of said target
materials, and
[Limitation 1(j)] to trigger an ejection system at a preset
time delay equal to the distance between said camera and said
ejection system divided by the rate of travel of the belt; and
[Limitation 1(k)] said ejection system disposed
immediately after said second end of said belt, whereby said
desired materials are separated from said heterogeneous
material,
[Limitation 1(l)] said ejection system being triggered by
said computer.

9. The system according to claim 1 wherein said
heterogeneous materials comprise waste and recyclable
materials.
Ex. 1001, 5:66–6:32, 6:58–59; Pet. 40–51.

IPR2020-00899
Patent 9,950,346 B2

7

D. References Relied Upon
Petitioner’s challenges rely on the following references (Pet. 4–6):
Name Reference Ex. No.
Carlsohn Excerpts from Color Image Processing:
Methods and Applications, CRC Press
(2007)
1002
Sinram WO Publication No. 2007/112591 A1,
published Oct. 11, 2007
1003
Arleth US Patent Publication No. 2004/0044436
A1, published Mar. 4, 2004
1004
Kampf US Patent Publication No. 2005/0127285
A1, published June 16, 2005
1005
Ulrichsen US Patent No. 6,353,197 B1, issued Mar.
5, 2002
1006
Mayhak US Patent No. 5,339,961, issued Aug. 23,
1994
1007
Cowling2 US Patent Publication No. 2007/0278139
A1, published Dec. 6, 2007
1019

IPR2020-00899
Patent 9,950,346 B2

8

E. Alleged Grounds of Unpatentability2
Petitioner contends that the challenged claims are unpatentable based
on the following asserted grounds (Pet. 4–6):
Ground Claim(s)
Challenged
35
U.S.C. §
Reference(s)/Basis3
1 1–5, 7, 9–12 103 Carlsohn, Arleth
2 8 103 Carlsohn, Arleth, Ulrichsen,
Mayhak
3 1–7, 9–16 103 Carlsohn, Arleth, Sinram
4 8 103

Carlsohn, Arleth, Ulrichsen,
Mayhak, Sinram
5 2 103 Carlsohn, Arleth, Kampf
6 2 103 Carlsohn, Arleth, Sinram,
Kampf
7 12, 13 103 Carlsohn, Arleth, Cowling2
8 12, 13 103 Carlsohn, Arleth, Sinram,
Cowling2
Petitioner supports its challenge with a declaration from Dr. Gonzalo
R. Arce (Ex. 1008).4 Patent Owner supports its Response with declarations
from Mr. Fred P. Smith (Exs. 2001, 2022).

2 The Leahy-Smith America Invents Act (“AIA”), Pub. L. No. 112-29, 125
Stat. 284, 287–88 (2011), amended 35 U.S.C. § 103, effective March 16,
2013. Because the application from which the ’346 patent issued asserts
priority to a parent application filed before this date, and this priority is not
at issue in this proceeding, we apply pre-AIA version of § 103.
3 Petitioner cites to the native page numbers of many of the exhibits. See,
e.g., Pet. 33 (citing to Carlsohn’s native page numbers 405–413, rather than
the corresponding exhibit’s page numbers 39–47). For clarity of record, our
IPR2020-00899
Patent 9,950,346 B2

9

II. ANALYSIS
A. Preliminary Response Incorporated by Reference
In the Response, Patent Owner incorporates by reference “all the
evidence and arguments previously presented by” Patent Owner, including,
we presume, those made in the Patent Owner’s Preliminary Response. See
PO Resp. 1–2.
We do not permit a Patent Owner to incorporate by reference
arguments made in the Preliminary Response. Our Rules provide that
“[a]rguments must not be incorporated by reference from one document into
another document.” 37 C.F.R. § 42.5(a)(3). Furthermore, our Scheduling
Order states that “Patent Owner is cautioned that any arguments not raised in
the response may be deemed waived.” Paper 12, 8 (bold emphasis omitted).
Accordingly, arguments made by Patent Owner in its Preliminary Response,
which were not otherwise made in the Response, are waived. See Tr. 4:23–
5:9 (explaining the same during the oral argument).

B. Level of Ordinary Skill in the Art
In determining whether an invention would have been obvious at the
time it was made, we consider the level of ordinary skill in the pertinent art
at the time of the invention. Graham v. John Deere Co. of Kansas City, 383
U.S. 1, 17 (1966).

citations will be to the exhibits’ page numbers, rather than the native page
numbers.
4 By e-mail communication from the Board on June 25, 2020, we authorized
Petitioner to file a corrected Exhibit 1008. Our reference to Exhibit 1008 is
to the corrected, second-filed exhibit, filed on July 7, 2020.
IPR2020-00899
Patent 9,950,346 B2

10

Factors pertinent to a determination of the level of ordinary skill in the
art include: (1) educational level of the inventor; (2) type of problems
encountered in the art; (3) prior art solutions to those problems; (4) rapidity
with which innovations are made; (5) sophistication of the technology; and
(6) educational level of workers active in the field. Environmental Designs,
Ltd. v. Union Oil Co., 713 F.2d 693, 696–97 (Fed. Cir. 1983) (citing
Orthopedic Equip. Co. v. All Orthopedic Appliances, Inc., 707 F.2d 1376,
1381–82 (Fed. Cir. 1983)). Not all such factors may be present in every
case, and one or more of these or other factors may predominate in a
particular case. Id. Moreover, these factors are not exhaustive but are
merely a guide to determining the level of ordinary skill in the art. Daiichi
Sankyo Co. Ltd, Inc. v. Apotex, Inc., 501 F.3d 1254, 1256 (Fed. Cir. 2007).
In determining a level of ordinary skill, we also may look to the prior
art, which may reflect an appropriate skill level. Okajima v. Bourdeau, 261
F.3d 1350, 1355 (Fed. Cir. 2001).
In the Institution Decision, we determined that a person having
ordinary skill in the art (“POSITA”) would have had a bachelor’s degree in
electrical engineering, computer science, computer engineering, or a related
subject, or four to five years of work experience with heterogeneous material
sorting systems used in materials recycling facilities. Inst. Dec. 12–13.
Neither Petitioner nor Patent Owner disputes our determination of the
level of ordinary skill in the art. See generally Pet. Reply; see also PO Resp.
5 (acknowledging our definition of a POSITA). Accordingly, we maintain
our definition of the level of ordinary skill in the art as set forth in our
Institution Decision. Inst. Dec. 12–13.

IPR2020-00899
Patent 9,950,346 B2

11

C. Claim Construction
In an inter partes review proceeding for a petition filed on or after
November 13, 2018, a patent claim shall be construed using the same claim
construction standard that would be used to construe the claim in a civil
action under 35 U.S.C. § 282(b). See 37 C.F.R. § 42.100(b) (2019). This
rule adopts the same claim construction standard used by Article III federal
courts (see id.), which follow Phillips v. AWH Corp., 415 F.3d 1303 (Fed.
Cir. 2005) (en banc) and its progeny. Under the Phillips standard, the words
of a claim are generally given their “ordinary and customary meaning,”
which is the meaning the term would have to a person of ordinary skill at the
time of the invention, in the context of the entire patent including the
specification. See Phillips, 415 F.3d at 1312–13.
Based on the final record before us, we determine that three clauses
within the claims require express construction for purposes of this Final
Decision. We address each in turn below.
1. “at least one hyperspectral camera said hyperspectral camera
configured to receive spectral data from a plurality of selected
spectral bands of infrared and visible light corresponding to
spectral signatures of target materials to be identified”
At the institution phase, we determined that the plain and ordinary
meaning of the claimed limitation, “said hyperspectral camera configured to
receive spectral data from a plurality of selected spectral bands of infrared
and visible light” as recited in Limitation 1(b), required a single
hyperspectral camera configured to receive spectral data from a plurality of
selected spectral bands of infrared (“IR”) and visible light. Inst. Dec. 14–17.
Petitioner does not dispute our construction (see generally Pet. Reply)
and Patent Owner expressly agrees with this construction (see PO Resp. 5
IPR2020-00899
Patent 9,950,346 B2

12

(“PO believes that construction is correct, as far as it goes.”)). Accordingly,
we maintain the construction set out in our Institution Decision. See Inst.
Dec. 14–17.
Patent Owner submits, however, that the claim language
“corresponding to spectral images of target materials to be identified”
requires that “the camera specifically be configured to receive spectral data
up to at least 1700 nm (nanometers) in order to receive spectral data
corresponding to spectral signatures of target materials to be identified, i.e.
recyclable materials and commodities.” Id. at 5–6 (citing Ex. 2022 ¶¶ 6–7)
(emphasis added); see also PO Sur-Reply 4 (“[T]he intrinsic evidence
discloses that the claimed hyperspectral camera must image infrared
wavelengths up to 1700 nm.”).
In support of Patent Owner’s proposed construction, Patent Owner
points out that dependent claim 9 “requires that the heterogeneous materials
to be sorted be comprised of waste and recyclable materials.” PO Resp. 6.
Patent Owner further cites the Specification, which states, “Even another
embodiment of the present invention provides such a method wherein
heterogeneous material comprises waste material and recyclable material in
combination.” Id. (citing Ex. 1001, 3:21–24).
In addition to citing claim 9 and the Specification, Patent Owner cites
to statements made by John Green (a co-inventor of the ’346 patent) during
prosecution, including:
Another important aspect of our invention to the recycling
industry is its ability to sort previously un-sortable materials.
Specifically, our development of hyperspectral technology has
enabled the sorting of: black plastics, which are important to auto
scrap and industrial recyclers; chlorine based plastics, such as
PVC and vinyl so that they may be removed from the waste
IPR2020-00899
Patent 9,950,346 B2

13

stream prior to their disposal at waste energy plants, extending
the life of the plants’ burn chambers and thereby raising the value
of the feed stock; arsenic laden pressure treated wood, as
differentiated from natural wood, so that it may be removed prior
to power generation.
. . .
Recycling markets further require the flexibility to easily
differentiate between those plastics with the highest market value
with and those with the lowest market value, such as those having
outer plastic or inner wraps. In post-consumer plastics recycling,
50% of all incoming containers are PET [or polyethylene
terephthalate] “water bottle” plastics. In case the PET “eye”, or
optical measuring device, went off line, a downstream optical
sensor that could easily switch from HDPE plastics to PET and
sort natural HDPE [or high-density polyethylene] plastics from
Colored HDPE plastics without picking whites “all other sorting
technologies are unable to differentiate White HDPE bottles
from Natural HDPE” was desirable. Markets demand th[is]
flexibility.
Id. at 7–8 (citing Ex. 2002, 262–63).
Patent Owner further submits that, during prosecution, Peter Mendre
(the other co-inventor of the ’346 patent) “describ[ed] the use of
hyperspectral imaging in connection with the claimed invention . . . [and] a
particular type of plastic sorting” that the inventors “were trying to achieve,
providing the below illustration:”
IPR2020-00899
Patent 9,950,346 B2

14

Id. at 9–10 (citing Ex. 2002, 49; Ex. 2003, 129); PO Sur-Reply 5 (citing
Ex. 2002, 49, Ex. 2003, 129).5 Patent Owner submits that the above
“illustration shows how a POSITA could differentiate PET (a recyclable

5 The color figure provided by Patent Owner does not exist on either page 49
of Exhibit 2002 or page 129 of Exhibit 2003. Rather, these exhibits show a
black-and-white version of the figure provided by Patent Owner.
IPR2020-00899
Patent 9,950,346 B2

15

plastic) from a waste plastic such as HDPE.” PO Resp. 10 (citing Ex. 1008
¶ 59); PO Sur-Reply 5 (arguing the same).
Based on claim 9, the Specification, and the prosecution history of the
’346 patent, Patent Owner proposes that “the ’346 patent makes clear and
the inventors made very clear during prosecution that their invention need[s]
to be able to sort recyclable plastic waste from waste plastic, and to do that
one needed to collect IR spectral data up to at least 1700 nm.” PO Resp. 10.
Patent Owner further cites to Carlsohn—one of the references cited by
Petitioner—in support of its proposed construction. See id. at 10–12. In
particular, Patent Owner cites to Carlsohn’s Figure 17.11 (id. at 10), which
we reproduce, below:

Figure 17.11 depicts “[r]eflectance spectra of ABS [or acrylonitrile
butadiene styrene], PVC-U [or polyvinyl chloride] (unplasticized), PS [or
polystyrene], PE-HD [or polyethylene] (high-density), PETP [or PET] and
PP [or polypropylene] in the NIR [or near-infrared] band.” Ex. 1002, 45.
IPR2020-00899
Patent 9,950,346 B2

16

Carlsohn discloses that “[s]trong absorption features can be seen around
1200 nm and 1400 nm.” Id.
Patent Owner submits that a “POSITA understood, as explained by
the inventors, that to be a useful system, IR spectral data needed to cover at
least wavelengths from 900 to 1700 nm.” PO Resp. 11.6
In its Sur-Reply, Patent Owner further cites to Petitioner’s expert,
Dr. Arce, for “repeatedly mention[ing] the wavelength ranges of the prior art
documents as allegedly reading on the claims of the ’346 patent because the
prior art included identification systems measuring reflectance of material
between 400 nm and 1750 nm, and specifically the NIR range of 950-1750
nm.” PO Sur-Reply 8–9 (citing Ex. 1008 ¶¶ 27, 36, 37, 42, 59, 66, 67).
In its Reply, Petitioner contends that “Patent Owner tries to assert that
the claims require at least one camera be able to cover wavelengths through
the visible range up to the infrared wavelength of 1700 nm. This, however,
is Patent Owner conjuring limitations to avoid the inevitable cancellation of
claims.” Pet. Reply 7. Petitioner submits that “Patent Owner offers no
instance in the specification, claims or prosecution history where the
inventor says that the claims cover any particular wavelengths in the infrared
range, let alone 1700 nm.” Id. (citing Ex. 1038, 28:3–22 (Smith
deposition)).
We agree with Petitioner.
Any special definition for a claim term must be set forth in the
specification with reasonable clarity, deliberateness, and precision. In re

6 Patent Owner does not cite to any testimony, including Mr. Smith’s, in
support of this assertion. See PO Resp. 11.
IPR2020-00899
Patent 9,950,346 B2

17

Paulsen, 30 F.3d 1475, 1480 (Fed. Cir. 1994). Here, no such special
definition exists.
As explained above, the words of a claim are generally given their
“ordinary and customary meaning,” which is the meaning the term would
have to a person of ordinary skill at the time of the invention, in the context
of the entire patent including the specification. See Phillips, 415 F.3d at
1312–13.
Combing through Patent Owner’s numerous pages of briefing, as well
as the Specification and the prosecution history, we find nothing in the
record to support Patent Owner’s proposed construction that the claimed
limitation requires a “single hyperspectral camera configured to receive
spectral data from a plurality of selected spectral bands of visible and IR
light up to at least 1700 nm in wavelength.” PO Resp. 12 (emphasis added);
see also PO Sur-Reply 10 (proposing the same). Petitioner is correct that the
“1700 nm target is a purely arbitrary value, picked out of the air.” Pet.
Reply 8; see also Tr. 8:23–12:1 (“Our point here is that this is truly
arbitrative pluck, 1,700 nanometers, out of the air.”).
Patent Owner argues that claim 1 requires a 1700 nm wavelength limit
based on a figure referenced by an inventor during prosecution (see PO
Resp. 9) and a figure from Carlsohn (see id. at 10). There are multiple
problems with Patent Owner’s construction, however.
First, nothing in the ’346 patent mentions any particular wavelengths,
let alone one that is 1700 nm. Indeed, Patent Owner’s expert and counsel
acknowledged as much. See Ex. 1038, 28:3–18 (Mr. Smith testifying on
cross examination that he does not “think there is a particular” wavelength
disclosed in the ’346 patent); see also Tr. 46:9–10 (“Specific wavelengths
IPR2020-00899
Patent 9,950,346 B2

18

are not mentioned in the specification.”). Moreover, Patent Owner appeared
to acknowledge that the Specification does not provide explicit written
description support for a hypothetical limitation requiring receiving spectral
data up to at least 1,700 nanometers. See Tr. 45:10–47:8 (“a POSITA
wouldn’t know that you need to go up to 1,700 to sort multiple recyclable
materials” but “it would have to come down to the language discussing --
that mentions recyclable materials, plural”).
Second, the plain and ordinary language of claim 1 simply recites
“heterogeneous material.” See Ex. 1001, 5:66–6:32. Although dependent
claim 9 further recites “waste and recyclable materials,” neither independent
claim 1 nor dependent claim 9 recites “plastic,” let alone polyethylene
terephthalate or high-density polyethylene shown in the figure cited by
Patent Owner. See id. at 5:66–6:32, 6:58–59; see also PO Resp. 9
(reproducing a figure referenced by an inventor during prosecution).
Furthermore, we find no reference to “plastic,” let alone HDPE or PET, in
the Specification. See Pet. Reply 8 (finding the same). Even if the claims
and Specification recited “HDPE” or “PET,” however, nothing in the claims
or Specification would lead us to believe that the material recited required
any particular wavelength for “corresponding to spectral signatures of” that
particular plastic, as Patent Owner argues. See Ex. 1001, 5:66–6:32.
Third, Patent Owner’s reliance on Carlsohn’s Figure 17.11 fails to
support Patent Owner’s assertion that a “POSITA understood, as explained
by the inventors, that to be a useful system, IR spectral data needed to cover
at least wavelengths from 900 to 1700 nm.” PO Resp. 11. We agree with
Petitioner that Carlsohn’s Figure 17.11 “shows that the relevant data trails
off at about 1400 nm, well before 1700 nm.” Pet. Reply 9. Indeed,
IPR2020-00899
Patent 9,950,346 B2

19

Carlsohn’s own description of Figure 17.11 refers to the “[s]trong absorption
features . . . seen around 1200 nm and 1400 nm,” reflecting Carlsohn’s focus
on wavelengths much less than 1700 nm. Ex. 1002, 45. We further find that
PETP—which is another name for PET—appears to drop off to zero at about
1650 nm, not 1700 nm. See Ex. 1002, 45 (Fig. 17.11); see also id. at 40
(defining polyethylene terephthalate as “PET” or “PETP”).
Fourth, Patent Owner’s reliance on Dr. Arce’s testimony is inapposite.
See PO Sur-Reply 8–9. Even if Dr. Arce believes that “the prior art
documents . . . read[] on the claims of the ’346 Patent because the prior art
included identification systems measuring the reflectance of material
between 400 nm and 1750 nm” (id. at 8), this does not mean that we must
construe the claim as requiring a 1700 nm wavelength range. In other
words, simply because wavelengths falling between 400 to 1750 nm may
satisfy the claimed limitation requiring a “plurality of selected spectral bands
of infrared and visible light,” this is not to say that the claim requires
receiving “spectral data in the visible and infrared ranges up to at least
1700 nm,” as Patent Owner proposes. See id. at 10.
Mr. Smith’s testimony that “[c]ameras utilizing InGaAs sensors (the
standard for recycling) generally work in the 1000 nm to 1700 nm range”
(Ex. 2022 ¶ 6) is not persuasive for similar reasons. In addition, this
testimony relies on a website article published in 2016, well after the
’346 patent’s 2008 priority date. See id.; Ex. 1001, code (60). Further, the
quotation from the website article provided by Mr. Smith fails to support his
proposition that InGaAs sensors are “the standard for recycling,” in any time
frame. See Ex. 2022 ¶ 6.
IPR2020-00899
Patent 9,950,346 B2

20

For at least these reasons, we agree with Petitioner that “there is no
reason to arbitrarily build in a 1700 nm limitation into a claim.” Pet. Reply
10. Accordingly, we do not construe “at least one hyperspectral camera said
hyperspectral camera configured to receive spectral data from a plurality of
selected spectral bands of infrared and visible light corresponding to spectral
signatures of target materials to be identified” as requiring a camera
“configured to receive spectral data up to at least 1700 nm,” as Patent Owner
proposes. See PO Resp. 6.

2. “said camera disposed proximate to a visible or infrared light
source”
At the institution phase, we construed “disposed proximate to” by its
plain and ordinary meaning, and determined that the plain and ordinary
meaning of “proximate” is “nearby. Inst. Dec. 18. Accordingly, we
construed “disposed proximate to a visible or infrared source,” as recited in
Limitation 1(c), to mean “disposed nearby” the visible or infrared source.
See id. We further note that during prosecution of the ’346 patent, the
applicant acknowledged that the term “‘proximate’ is broad in its scope.”
See Ex. 1020, 70 (emphasis added).
Petitioner does not dispute our construction (see generally Pet. Reply)
and Patent Owner expressly agrees with this construction (see PO Resp. 12
(“PO does not here challenge that construction.”)). Accordingly, we
maintain the construction set out in our Institution Decision. See Inst. Dec.
18.

IPR2020-00899
Patent 9,950,346 B2

21

3. “said hyperspectral identification system being disposed over
said second end of said belt”
Patent Owner proposes that “said second end of said belt” should be
construed to mean “the portion of the belt immediately adjacent the ejection
system.” PO Resp. 14 (emphasis added); see also id. at 22. As such, Patent
Owner explains that the claimed limitation, “said hyperspectral identification
system being disposed over said second end of said belt,” as recited in
Limitation 1(f), should be construed to mean “the hyperspectral
identification system is positioned over the portion of the belt immediately
adjacent the ejection system.” Id. at 14 (emphasis added).
Patent Owner explains that the “figures of the ’346 Patent support this
understanding,” referencing Figures 1 and 5. Id. In addition to Figures 1
and 5, Patent Owner cites to the text of the Specification (see id. at 16–17)
and the prosecution history (see id. at 17).
Petitioner’s position is that “a POSITA would understand the term
[end] to be a term of approximation.” See Pet. Reply 15 (citing Whirlpool
Corp. v. LG Elecs., Inc., No. 1:04-CV-100, 2005 U.S. Dist. LEXIS 44479
(W.D. Mich. Nov. 8, 2005)). Petitioner explains that “[t]he term ‘second
end’ must have been meant to convey a general location, as opposed to a
particular point on the belt; otherwise, the system would not have time to
process the data before items contacted the ejector system, rending the
system inoperable for its intended use.” Id.
We agree with Petitioner.
We find nothing cited by the Patent Owner in the Specification or
prosecution history to support Patent Owner’s narrow construction that
“second end” must be construed as “immediately adjacent the ejection
system.” To illustrate, we address Figures 1 and 5 of the ’346 patent, which
IPR2020-00899
Patent 9,950,346 B2

22

Patent Owner directs our attention to. See PO Resp. 14–16. We first
reproduce Figure 1 of the ’346 patent, below:

Figure 1 “is a front perspective drawing illustrating a system for the
separation of heterogeneous material configured in accordance with one
embodiment of the present invention.” Ex. 1001, 1:36–38.
We further reproduce Figure 5 of the ’346 patent, below:
IPR2020-00899
Patent 9,950,346 B2

23

Figure 5 “is a front perspective drawing illustrating a conveyor [1], air knife
and ejector [2] configured for use in a system for the separation of
heterogeneous material configured in accordance with one embodiment of
the present invention.” Ex. 1001, 1:50–57, 5:1–2, 5:11, Fig. 6 (exploded
view of air knife and ejector 2).
Patent Owner submits that “hyperspectral identification system, item
4, is disposed over the second end of the belt, which is the portion of the belt
(1) immediately adjacent the ejection system (2).” PO Resp. 16. Patent
Owner quotes the following descriptions from the Specification:
IPR2020-00899
Patent 9,950,346 B2

24

(1) “Immediately after the discharge of the accelerator
conveyor 1 is disposed an air knife 2 that helps to levitate
materials and decreases the likelihood of material lodging
between the accelerator belt 1 and manifold ejector plate 14.” PO
Resp. 16–17 (quoting Ex. 1001, 5:9–13); and
(2) “Unselected items typically fall off the end of the optics
conveyor or shaker deck and [are] dropped on a conveyor below
that would transport that material to another optical system
material where the process occurs again.” PO Resp. 17 (quoting
Ex. 1001, 1:41–45) (emphasis omitted).
We do not find Figures 1 and 5 of the ’346 patent, the cited portions
of the Specification, or the prosecution history as requiring “said second end
of said belt” to mean “the portion of the belt immediately adjacent the
ejection system,” as argued by Patent Owner. See PO Resp. 14 (emphasis
added). Any special definition for a claim term must be set forth in the
specification with reasonable clarity, deliberateness, and precision. Paulsen,
30 F.3d at 1480. Here, even if Figures 1 and 5 depict a hyperspectral
identification system 4 is disposed over and immediately adjacent the
terminus of the second end of the belt, we do not see the Specification as
providing a special definition or meaning of the claimed limitation that
requires “said second end of said belt” to mean “the portion of the belt
immediately adjacent the ejection system” (see id. at 14). Without such a
special definition or meaning, limitations may not be read from
the specification into the claims. See In re Van Geuns, 988 F.2d 1181, 1184
(Fed. Cir. 1993). Rather, we agree with Petitioner that the term “second end
of said belt” is a term of approximation. See Pet. Reply 15.
Furthermore, we agree with Petitioner’s explanation that construing
the claim to require the camera to be positioned immediately adjacent, or at
the terminus of the belt, would render superfluous the claim 1 limitation,
IPR2020-00899
Patent 9,950,346 B2

25

“trigger[ing] an ejection system at a preset time delay equal to the distance
between said camera and said ejection system divided by a rate of travel of
the belt.” See Pet. Reply 14 (alteration in original); see Stumbo v. Eastman
Outdoors, Inc., 508 F.3d 1358, 1362 (Fed. Cir. 2007) (explaining that
a claim interpretation that renders a claim term or phrase superfluous is
disfavored).
Claim 1 requires that the “ejection system [is] disposed immediately
after said second end of said belt,” and we do not see how there can be a
“time delay equal to the distance between said camera and said ejection
system divided by a rate of travel of the belt” if we were to adopt Patent
Owner’s construction of “second end” as limited to the belt terminus. See
Ex. 1001, 6:18–20, 6:29–30. In other words, requiring the camera to be
positioned “over” the belt terminus per Patent Owner’s construction
(Ex. 1001, 6:17–20), and requiring the ejection system to be immediately
after the belt terminus (id. at 6:29–30) would leave no room for a “distance
between said camera and said ejection system,” as the “distance between
said camera and said ejection system” would be zero. See Pet. Reply 14
(explaining the same); see also Ex. 1001, 6:25–28. The camera and the
ejection system would both be located at the belt terminus.
Patent Owner’s expert appears to agree with our understanding, as
well. In particular, Mr. Smith acknowledged that there is a “distance
between where the camera picks up the image and the ejection system.” See
Ex. 1038, 58:14–19 (“Q Well, the point on the ejection system when it first
makes contact with the – with the material? A It’s actually the distance
between where the camera picks up the image and the ejection system.”); see
also Pet. Reply 14 (explaining the same). As to whether the camera or
IPR2020-00899
Patent 9,950,346 B2

26

identification system is “immediately adjacent” the second end of the belt,
Mr. Smith “totally agree[d]” that the camera cannot be at the “absolute edge
of the second belt.” See Ex. 1038, 60:17–21 (“Q When we talk about the
camera being over the second end of the belt, we have to assume, as a
POSITA, that it’s not at the absolute edge of the second belt, don’t we? A I
totally agree.”); see also id. at 61:3–6 (“Q And the claim presumes that the
imaging system would not be disposed over the absolute edge on the second
end of the belt? A Yeah.”).
Accordingly, we do not construe “said hyperspectral identification
system being disposed over said second end of said belt” to mean “the
hyperspectral identification system is positioned over the portion of the belt
immediately adjacent the ejection system,” as proposed by Patent Owner.
See PO Resp. 14 (emphasis added). Rather, we agree with Petitioner that the
term “second end” is a term of approximation, and “convey[s] a general
location, as opposed to a particular point on the belt” so that the
hyperspectral identification system would “have the time to process the data
before items contacted the ejector system.” See Pet. Reply 15.

4. Other Claim Terms
There are no other terms that require express construction for
purposes of this Decision. See Nidec Motor Corp. v. Zhongshan Broad
Ocean Motor Co., 868 F.3d 1013, 1017 (Fed. Cir. 2017) (“we need only
construe terms ‘that are in controversy, and only to the extent necessary to
resolve the controversy’”) (quoting Vivid Techs., Inc. v. Am. Sci. & Eng’g,
Inc., 200 F.3d 795, 803 (Fed. Cir. 1999)).

IPR2020-00899
Patent 9,950,346 B2

27

D. Prior Art Status of Carlsohn
Before reaching the merits of Petitioner’s contentions, all of which
rely on Carlsohn (see Pet. 4–6), we must determine as a threshold matter
whether Carlsohn is a prior art printed publication under 35 U.S.C. § 102(b).
Carlsohn is a sixty-one page excerpt from a book titled “Color Image
Processing Methods and Applications” (“the Book”). Ex. 1002, 1.
Petitioner relies on Chapter 17 of the Book (see, e.g., Pet. 33), which is titled
“Spectral Imaging and Applications” (Ex. 1002, 27).
At the institution stage, and for purposes of institution, we determined
that “Petitioner identifies, with particularity, evidence to establish a
reasonable likelihood that Carlsohn was publicly accessible before the
critical date of the challenged patent.” Inst. Dec. 22; see also id. at 19–22
(addressing Patent Owner’s arguments that Carlsohn is not a printed
publication).
In its Response, Patent Owner argues that Carlsohn is not a printed
publication under 35 U.S.C. § 102(b) and that Petitioner failed to meet its
burden in establishing that Carlsohn qualifies as a printed publication. See
PO Resp. 24–35. The parties further address this issue in Petitioner’s Reply
(Pet. Reply 1–6) and Patent Owner’s Sur-Reply (PO Sur-Reply 1–4).

1. Relevant Law
Petitioner bears the burden of establishing by a preponderance of the
evidence that Carlsohn is a printed publication. See 35 U.S.C. § 316(e);
Nobel Biocare Servs. AG v. Instradent USA, Inc., 903 F.3d 1365, 1375 (Fed.
Cir. 2018); Hulu, LLC v. Sound View Innovations, LLC, IPR2018-01039,
IPR2020-00899
Patent 9,950,346 B2

28

Paper 29 at 11 (PTAB Dec. 20, 2019) (precedential) (citing Nobel Biocare,
903 F.3d at 1375).
The determination of whether a document is a “printed publication”
under 35 U.S.C. § 102 “involves a case-by-case inquiry into the facts and
circumstances surrounding the reference’s disclosure to members of the
public.” In re Klopfenstein, 380 F.3d 1345, 1350 (Fed. Cir. 2004).
“Because there are many ways in which a reference may be disseminated to
the interested public, ‘public accessibility’ has been called the touchstone in
determining whether a reference constitutes a ‘printed publication’.” Jazz
Pharm., Inc. v. Amneal Pharm., LLC, 895 F.3d 1347, 1355 (Fed. Cir. 2018)
(quoting In re Hall, 781 F.2d 897, 898–99 (Fed. Cir. 1986)). A reference is
considered publicly accessible if it was “disseminated or otherwise made
available to the extent that persons interested and ordinarily skilled in the
subject matter or art, exercising reasonable diligence, can locate it.’” Id. at
1355–56 (quoting In re Wyer, 655 F.2d 221, 226 (CCPA 1981)); see also
Medtronic, Inc. v. Barry, 891 F.3d 1368, 1380 (Fed. Cir. 2018) (“reference
will be considered publicly accessible if it was ‘disseminated or otherwise
made available to the extent that persons interested and ordinarily skilled in
the subject matter or art exercising reasonable diligence[] can locate it’”)
(internal quotation omitted).

2. Petitioner’s Evidence and Position
Petitioner asserts that “Carlsohn was published over one year before
November 18, 2008, the ’346 Patent’s earliest claimed priority date, and thus
is prior art under § 102(b),” citing the sworn testimony of Dr. Arce. Pet. 4
(citing in part Ex. 1008 ¶¶ 52–56). In support of this assertion, Petitioner
IPR2020-00899
Patent 9,950,346 B2

29

cites to Carlsohn’s 2007 copyright date (id. (citing Ex. 1002, 1, 5–10)) and
submits a May 2007 Amazon.com webpage that appears to show the Book
(containing Carlsohn) for sale before the critical date (see id. at 4–5 (citing
Ex. 1032)). Petitioner further submits two webpage screenshots (Exs. 1033,
1034) that identify the Book as a “New Book Published!” before the critical
date. See Pet. 5. Petitioner also submits evidence that over 550 copies of
the Book had been sold to the public prior to the critical date. Pet. Reply 1
(citing Ex. 1037 (“Ms. Militello’s testimony”)).

3. Analysis
Patent Owner’s argument attacks generally the veracity and
sufficiency of Petitioner’s evidence. See PO Resp. 24–35; see also PO Sur-
Reply 1–4.
After considering the arguments and weighing the evidence presented
by the parties, we credit Dr. Arce’s testimony and Ms. Militello’s testimony
and find that Carlsohn was disseminated or otherwise made available to the
extent that POSITAs, exercising reasonable diligence, could locate it before
the critical date of the invention. See Medtronic, 891 F.3d at 1380. In
particular, we find that Petitioner’s evidence and testimony that over 550
copies of the Book had been sold prior to the critical date establishes, on its
own, that Carlsohn is a prior art printed publication under 35 U.S.C.
§ 102(b). We further find that Petitioner’s evidence and testimony that the
Book was available for sale on the Internet before the critical date of the
invention also establishes, on its own, that Carlsohn is a prior art printed
publication.
We address each of Patent Owner’s arguments separately, below.
IPR2020-00899
Patent 9,950,346 B2

30

First, Patent Owner argues that “neither Petitioner nor its declarant,
Dr. Arce, explain how, why, or when a POSITA could have located
Carlsohn or would have bothered to review Carlsohn to find the relevant
section here relied upon, let alone provide competent evidence establishing
either proposition.” PO Resp. 25. Patent Owner further argues that neither
Petitioner nor Dr. Arce attempt to establish that the particular discussion
relied upon by Petitioner—namely, Chapter 17’s SpectroSort discussion—
was “uploaded to a website, deposited in an indexing library, or in some
other way meaningfully indexed prior to November 18, 2007.” Id. at 26–27;
see also PO Sur-Reply 2 (“Finding a chapter in a compilation is tantamount
to finding a needle in a haystack”). Patent Owner asserts that “neither
provides the Board with any explanation on how the allegedly relevant
information found on pages 407-412 of [Carlsohn] could have been located
by a member of the public at any point in time.” PO Resp. 27. Patent
Owner also argues that “Petitioner must show that a POSITA could
meaningfully, exercising reasonable diligence, find Carlsohn, a chapter in
the Book.” Id. Patent Owner contends that “even if Dr. Arce had conducted
the reasonable diligence of a POSITA, the Amazon Listing would not have
been found” and that “neither Petitioner nor Dr. Arce show how a POSITA
would even locate [Carlsohn] out of the millions of listings on
Amazon.com.” Id. at 29; see also id. at 30 (“Petitioner has failed to explain
how the Amazon Listing’s mere existence constitutes accessibility.”).
Patent Owner further attacks Petitioner’s expert, Dr. Arce, for failing
to correctly identify search terms that a user would have used to locate
Carlsohn, and contends that “even if Dr. Arce had conducted the reasonable
diligence of a POSITA, the Amazon listing would not have been found.”
IPR2020-00899
Patent 9,950,346 B2

31

See id. at 29; see also id. (“[W]hen asked what search terms a POSITA
might use, the search terms Dr. Arce provided do not appear anywhere on
the Amazon listing.”).
Through the supplemental declaration of Mr. Smith (Ex. 2022), Patent
Owner submits the following evidence to support its position: “Documents
on Color Image Processing that Don’t Have Anything on Hyperspectral
Imaging” (Exhibit E of Exhibit 2022); Google Search Pages (Exhibit F of
Exhibit 2022); and Yahoo Search Pages (Exhibit G of Exhibit 2022). Mr.
Smith testifies that these exhibits “show[] how unlikely a POSITA would be
to search further into a book entitled ‘Color Image Processing’ for good
information on hyperspectral imaging in the NIR range or for help with
recycling in general.” Ex. 2022 ¶ 12 (addressing Exhibit E); id. ¶¶ 15–17
(addressing Exhibit F); id. ¶¶ 18–19 (addressing Exhibit G).7
We do not agree with Patent Owner.
Patent Owner’s contentions that a POSITA would not have been able
to locate the particular pages within Chapter 17 of the Book (i.e., Carlsohn)
is premised on an overly narrow test for determining public accessibility.
The test does not require Petitioner to identify the particular search terms
used for locating the Book. Further, the test for determining public
accessibility does not require a specific showing that sets forth the particular
search terms that a skilled artisan would have used to locate the particular
pages of a multi-chapter publication. See Pet. Reply 3 (arguing the same).
As explained above, determination of whether a document is a “printed

7 Patent Owner’s filing of Exhibits E, F, and G as part of and combined with
Exhibit 2022 is prohibited under our Rules. 37 C.F.R. § 42.6(a)(3). We
nonetheless consider these documents here.
IPR2020-00899
Patent 9,950,346 B2

32

publication” under 35 U.S.C. § 102 “involves a case-by-case inquiry into the
facts and circumstances surrounding the reference’s disclosure to members
of the public.” Klopfenstein, 380 F.3d at 1350 (emphasis added). “Because
there are many ways in which a reference may be disseminated to the
interested public, ‘public accessibility’ has been called the touchstone in
determining whether a reference constitutes a ‘printed publication.’” Jazz
Pharm., 895 F.3d at 1355 (emphasis added).
Having conducted a “case-by-case inquiry into the facts and
circumstances,” as set forth in Klopfenstein, and following the guidance of
Jazz Pharm. that “there are many ways in which a reference may be
disseminated to the interested public,” we find that Petitioner has shown, for
the reasons explained below, that Carlsohn was publicly accessible and
disseminated and qualifies as a prior art printed publication under 35 U.S.C.
§ 102(b).
As to Patent Owner’s evidence that a POSITA would not have found
Carlsohn based on particular terms used with Google and Yahoo search
engines (see Ex. 2022 ¶¶ 15–21, Exs. F, G), missing from Patent Owner’s
evidence is a similar search performed on Amazon, where Petitioner submits
evidence that Carlsohn was sold before the critical date (see Ex. 1032).
Patent Owner’s Google and Yahoo evidence does not directly address
Petitioner’s evidence that Carlsohn was publicly accessible through
Amazon.
Patent Owner further argues that the main title of the Book, “Color
Image Processing,” “is so general a term of art that countless applications
could be relevant.” PO Resp. 30–31. As to the editor of the Book, Patent
Owner contends that “Dr. Arce does not explain how a POSITA would even
IPR2020-00899
Patent 9,950,346 B2

33

be aware of the editor, Rastislav Lukac, testifying that he ‘[doesn’t] know if
somebody in the field would know him or not.’” Id. at 31 (quoting
Ex. 2023, 84:5–20 (Acre deposition)). As to the publisher of the Book,
Patent Owner argues that Petitioner “fail[s] to show accessibility because
[Exhibit 1037] does not establish that a POSITA would have known CRC
Press to be an established publisher in heterogeneous material sorting
systems.” Id. at 32. Patent Owner explains that “CRC Press’ parent
company, Taylor and Francis Group, is purportedly in the business of
publishing academic texts . . . [, however,] a POSITA may not have a
college degree, with his/her sole relevant experience being in the field of
heterogeneous material sorting systems used in materials recycling
facilities.” Id. (citing Ex. 1037 ¶ 2; Inst. Dec. 12–13 (defining POSITA)).
Patent Owner further argues, “Missing from Petitioner’s evidence is what
would have led a POSITA, i.e., someone interested in heterogeneous waste
separation using a hyperspectral imaging system, to seek out the Book prior
to November 18, 2007.” Id. at 33 (citing Ex. 2001 ¶¶ 44–46).
As to these arguments, we also find them unavailing. Even if we take
as true Patent Owner’s assertions that CRC Press and Taylor and Francis
Group are in the business of publishing academic texts, and that these
publishers, along with Editor Rastislav Lukac, are not well-known within the
heterogeneous material sorting systems community, Patent Owner’s
argument again applies an overly narrow test for public accessibility. See
Pet. Reply 3 (arguing the same). There is no requirement that the publisher
or its editor be well known within the particular field that a POSITA
operates within. Moreover, there is no requirement that the title of the book
IPR2020-00899
Patent 9,950,346 B2

34

be uniquely descriptive to the subject matter searched, in order to aid a
skilled artisan in finding the particular subject matter of his or her interest.
After weighing the testimony of Petitioner’s and Patent Owner’s
declarants, we credit Ms. Militello’s testimony (Ex. 1037) and find that it
establishes that Carlsohn was publicly accessible and, in fact, disseminated,
before the critical date of the invention.
Exhibit 1037 is a declaration from Corinne Militello, the Vice
President and Assistant General Counsel for Taylor & Francis Group
(Americas). Ex. 1037 ¶¶ 1–2. Ms. Militello testifies that she was “asked to
verify certain details about the publication of Color Image Processing:
Methods and Applications, CRC Press © 2007.” Id. ¶ 3. Ms. Militello
further testifies that
This publication date of October 20, 2006 is when CRC
Press made this book available to purchasers and third parties
such as universities, teachers, libraries, and resellers. From
October 20, 2006 to November 17, 2007, CRC Press advertised
this book, and shipped numerous physical copies of this book, to
such third parties. The confidential internal business records of
CRC Press indicate that CRC publically distributed over 550
copies of Color Image Processing: Methods and Applications
prior to November 17, 2007, mostly (over 500 copies) as actual
sales. CRC continued to advertise and ship this book from
November 17, 2007 to November 17, 2008, and thereafter.
Id. ¶ 5.
Specifically, we credit Ms. Militello’s testimony, including that “CRC
publically distributed over 550 copies of Color Image Processing: Methods
and Applications prior to November 17, 2007, mostly (over 500 copies) as
actual sales.” Id. We find that her testimony alone establishes that the Book
including Carlsohn was published on October 20, 2006, and was publicly
accessible before the critical date of the invention, November 17, 2007.
IPR2020-00899
Patent 9,950,346 B2

35

In addition to Ms. Militello’s testimony, we further find persuasive
Petitioner’s Amazon sales evidence (Ex. 1032) as evidence that the Book
and its Carlson chapter was publicly accessible before the critical date. This
evidence further supports a finding that Carlsohn was a printed publication
before the critical date of the invention. The webpage from Amazon.com
shows the Book (containing Carlsohn) for sale on May 9, 2007 (Ex. 1032,
1–2 (header information)), which is before the critical date of November 17,
2007. The book for sale on the Amazon webpage matches Exhibit 2002
(Carlsohn), as both identify the same title (“Color Image Processing
Methods and Applications”), cover (depicting yellow, orange, blue, and pink
flowers over a black background with binary numbers), ISBN (0-8493-9774-
X), editors (Rastislav Lukac and Konstantinos N. Plataniotis), and publisher
(CRC Press). Compare Ex. 1032, 1, with Ex. 1002, 1–5. We further note
that the ISBN number from Carlsohn also matches that of Ms. Militello’s
testimony. Compare Ex. 1002, 5 (“ISBN (0-8493-9774-X)), with Ex. 1037
¶¶ 3 (“As shown on the publication page, this book bears the International
Standard Book Number (ISBN)-10 of 0-8493-9774-X, and the ISBN-13 of
978-0-8493-9774-5”), 4 (“this webpage corresponds to the book bearing
ISBN 9780849397745, which omits the hyphens but is the same as the
ISBN-13 reflected on the publication page of Appendix A”).
We also find persuasive the testimony of Petitioner’s expert, Dr. Arce,
who testifies that Carlsohn is prior art under 35 U.S.C. § 102(b). See
Ex. 1008 ¶¶ 52–56. Dr. Arce testifies that the Amazon.com web listing
(Ex. 1032) “is a May 9, 2007 Wayback Machine printout of an [Amazon]
webpage.” Id. ¶ 53. Dr. Arce further states that the two webpage printouts
that identify Carlsohn as a “New Book Published!” are also Wayback
IPR2020-00899
Patent 9,950,346 B2

36

Machine printouts that identify Carlsohn as available for purchase in
October 2006 (Ex. 1033) and July 2007 (Ex. 1034). See Ex. 1008 ¶¶ 54–56.
Finally, Patent Owner argues that Exhibits 1032–1034 are “legally
deficient” as the “webpage exhibits have not been proven to be authentic . . .
and each is hearsay to which no evidentiary exception applies.” See PO
Resp. 34–35 (citing Fed. R. Evid. 803, 901). Although Patent Owner did not
file a motion to exclude Exhibits 1032–1034, Patent Owner submits that
“neither the webpage Exhibits, Exs. 1032-1034, nor Dr. Arce’s testimony
concerning them should be considered.” Id. at 35 (citation omitted).
In its Reply, Petitioner points out that Exhibits 1032, 1033, and 1034
are “located through the Wayback Machine . . . [and] each show[s] that
[Carlsohn] was available via Amazon and through the publisher’s website
prior to the critical date.” Pet. Reply 2 (citing Ex. 1008 ¶ 53). Dr. Arce
testifies that Exhibits 1032–1034 are Wayback Machine printouts from
www.amazon.com, www.colorimageprocessing.com/books, and
www.colorimageproceessing.org, respectively. See Ex. 1008 ¶¶ 53–55.
Dr. Arce, however, is a professor from the University of Delaware, not a
custodian from the Wayback Machine Internet Archive. See id. ¶¶ 4–5. In
addition to Dr. Arce’s testimony, however, Petitioner also cites to a
declaration from the Book’s publisher, and submits that “the publisher’s
declaration confirms that the book was published on October 20, 2006, and
that between then and November 17, 2007, it publicly distributed more than
550 copies of the book.” Pet. Reply 3 (citing Ex. 1037).
Patent Owner’s evidentiary objection to Exhibits 1032–1034 has some
merit. As explained above, Petitioner’s Exhibit 1032 is a Wayback Machine
screenshot of the Book for sale on amazon.com on May 9, 2007, and
IPR2020-00899
Patent 9,950,346 B2

37

included a hyperlink for interested viewers to “Search inside this book”;
Exhibit 1033 is a Wayback Machine screenshot from the publisher, CRC,
indicating the Book as a “New Book Published!” on October 25, 2006; and
Exhibit 1034 is Wayback Machine screenshot of a webpage created by the
book’s editors/authors, Rastislav Lukac and Kostas Plataniotis, indicating
the Book as sold on July 1, 2007. See Pet. Reply 2–3 (confirming the same
information). Although Petitioner relies on these exhibits as evidence that
Carlsohn was for sale and published before the critical date of the ’346
patent, Petitioner does not submit sworn testimony from a custodian of
records at the Wayback Machine Internet Archive to confirm the publication
dates of these screen shots. Despite this shortcoming, however, Patent
Owner’s argument ignores Petitioner’s other corroborating evidence,
namely, the sworn testimony of Ms. Militello. See PO Resp. 33–35. Taken
as a whole, including Ms. Militello’s uncontroverted testimony (Ex. 1037),
we find that the preponderance of the evidence supports a finding that
Carlsohn was publicly accessible before the critical date of the ’346 patent.
Importantly, Ms. Militello testifies that over 550 copies of Carlsohn were
sold and or distributed before the critical date of the ’346 patent. See
Ex. 1037 ¶ 5.

4. Summary
After considering the arguments and weighing the evidence presented
by the parties, we find that Petitioner has met its burden to show that
Carlsohn was disseminated or otherwise made available to the extent that
POSITAs, exercising reasonable diligence, could locate it before the critical
date of the invention. See Medtronic, 891 F.3d at 1380.
IPR2020-00899
Patent 9,950,346 B2

38

E. Principles of Law
“In an . . . [inter partes review], the petitioner has the burden from the
onset to show with particularity why the patent it challenges is
unpatentable.” Harmonic Inc. v. Avid Tech., Inc., 815 F.3d 1356, 1363 (Fed.
Cir. 2016). This burden never shifts to Patent Owner. Dynamic Drinkware,
800 F.3d at 1378.
Petitioner’s challenges are based on obviousness. Pet. 4–6; supra
§ I.E. A claim is unpatentable as obvious under 35 U.S.C. § 103(a) if the
differences between the claimed subject matter and the prior art are such that
the subject matter, as a whole, would have been obvious at the time the
invention was made to a person having ordinary skill in the art to which said
subject matter pertains. KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 406
(2007). The question of obviousness is resolved on the basis of underlying
factual determinations, including: (1) the scope and content of the prior art;
(2) any differences between the claimed subject matter and the prior art;
(3) the level of skill in the art; and (4) where in evidence, so-called
secondary considerations.8 Graham, 383 U.S. at 17–18.

F. Ground 1 – Claims 1–5, 7, and 9–12 Obvious over Carlsohn and Arleth
Petitioner asserts that claims 1–5, 7, and 9–12 are unpatentable as
obvious over Carlsohn and Arleth. Pet. 33.

8 Patent Owner has not submitted argument and objective indicia of
nonobviousness. See generally PO Resp.
IPR2020-00899
Patent 9,950,346 B2

39

1. Carlsohn (Ex. 1002)
Carlsohn is a sixty-one page excerpt from a book titled “Color Image
Processing Methods and Applications.” Ex. 1002, 1.9 Petitioner relies on
Chapter 17 of Carlsohn (see, e.g., Pet. 33), which is titled “Spectral Imaging
and Applications” (Ex. 1002, 27).
Carlsohn discloses a sorting system named SpectroSort that generates
hyperspectral images of heterogeneous waste to determine the identity and
location of the waste for subsequent sorting. See Ex. 1002, 39–47. To
illustrate SpectroSort, we reproduce Carlsohn’s Figure 17.10, below:

9 For clarity of record, our citations to Carlsohn will be to the exhibit’s page
numbers, rather than the book’s native page numbers. See supra n.2.
IPR2020-00899
Patent 9,950,346 B2

40

Figure 17.10 depicts “The SpectroSort© hardware setup” as “Reproduced
from the SpectroSort® product sheet.” Ex. 1002, 43. As shown in the
above figure, the SpectroSort includes a conveyor belt, an NIR/VIS sensor
system, and an air-jet array to sort products onto two different transport
belts. See id. at 41–43.

2. Arleth (Ex. 1004)
Arleth is a U.S. Patent Application titled “Process and Installation for
Separating Plastics, According to Type, From Recycled Plastic Mixtures.”
Ex. 1004, code (54). Arleth discloses a “process by which plastics are
separated from recycled plastic mixtures according to type.” Id. ¶ 1. Arleth
further discloses, “[t]he plastic parts to be selected and their position are
determined by optical sensors and [] sorting logic (stored program control
unit) in the optomechanical sorting device.” Id. To illustrate an
embodiment of Arleth’s system, we reproduce its Figure 4, below:

IPR2020-00899
Patent 9,950,346 B2

41

Figure 4 is a “schematic representation of a sorting system with two
subsequently arranged sorting devices with the same features.” Ex. 1004,
¶ 27.

3. Independent Claim 1
We address the claim limitations using Petitioner’s nomenclature as
identified above. See supra § I.C. In addressing the majority of the
limitations, Petitioner submits an annotated version of Carlsohn’s Figure
17.10 (Pet. 34), which we reproduce, below:

According to Petitioner, and as shown above, Figure 17.10 depicts a
hyperspectral camera that can capture a hyperspectral image, or
IPR2020-00899
Patent 9,950,346 B2

42

“hypercube,” of the waste by collecting both spatial and spectral
information. Id. Petitioner adds color to Carlsohn’s original Figure 17.10
and references this annotated figure in addressing several of the claim
limitations. See, e.g., id. at 41 (referencing the green “spectral imaging”
system).

a) Limitation 1(p) - A system for the identification and
sorting of heterogeneous material
Petitioner submits that Carlsohn’s SpectroSort system is a system for
sorting heterogeneous material, namely, different papers and polymers. Pet.
40 (citing Ex. 1002, 39–40, 44–46; Ex. 1008 ¶¶ 57–59, 70, 71).
Carlsohn discloses, “The main focus of SpectroSort is the sorting of
paper and polymers in an industrial environment.” Ex. 1002, 40. Carlsohn
explains that there are two kinds of paper that have to be separated:
(1) paper that can be freed of ink, or “de-inking” paper, such as white office
paper; and (2) “non-de-inking” paper, such as nonprinted cardboard. Id.
Carlsohn also explains that there are many different polymers to be sorted
for recycling, including polyethylene and polystyrene. Id.
Patent Owner does not dispute Petitioner’s assertions. See generally
PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitation 1(p), to the extent that the preamble is
limiting, for the reasons stated by Petitioner in the Petition, which we adopt
as our own findings, as further supported by the testimony of Dr. Arce.

IPR2020-00899
Patent 9,950,346 B2

43

b) Limitation 1(a) - a hyperspectral identification system
for capturing spectra of material
Petitioner asserts that Carlsohn discloses this limitation. Pet. 40–41
(citing Ex. 1002, 41–46; Ex. 1008 ¶¶ 61–64). Petitioner references
annotated Figure 17.10 (reproduced supra), and submits that the
“illumination unit” (annotated to be purple) illuminates the material to be
identified, so a spectral imaging unit (annotated to be green) receives the
reflected light. Pet. 41 (citing Ex. 1002, 41–43).
Carlsohn discloses, “The heart of SpectroSort® is a [near-infrared
(“NIR”)] spectral imaging system that is able to distinguish de-inking from
non-de-inking material based on those differences.” Ex. 1002, 41.
Dr. Arce testifies that SpectroSort’s spectral imaging system “includes
at least one hyperspectral camera configured to capture and (at least
temporarily) store hyperspectral images of waste that include both spatial
and spectral data: the visible . . . and near-infrared . . . sensor units.” Ex.
1008 ¶ 61; see also supra (annotated Figure 17.10).
Patent Owner does not dispute Petitioner’s assertions. See generally
PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitation 1(a) for the reasons stated by Petitioner
in the Petition, which we adopt as our own findings, as further supported by
the testimony of Dr. Arce.

IPR2020-00899
Patent 9,950,346 B2

44

c) Limitation 1(b) - said hyperspectral identification
system comprises at least one hyperspectral camera
said hyperspectral camera configured to receive
spectral data from a plurality of selected spectral
bands of infrared and visible light corresponding to
spectral signatures of target materials to be identified
if present in said heterogeneous material and spacial
data locating a position of said heterogeneous
materials on a solid belt10
As construed above, the claimed limitation “said hyperspectral camera
configured to receive spectral data from a plurality of selected spectral bands
of infrared and visible light” requires a single hyperspectral camera
configured to receive spectral data from a plurality of selected spectral bands
of infrared and visible light. See supra § II.C.1. In our claim construction,
we also determined that the limitation does not require the camera be
“configured to receive spectral data up to at least 1700 nm,” as Patent Owner
proposes. See id.; see also PO Resp. 6.
Petitioner submits that it would have been obvious to use ImSpector
V10’s spectral range of approximately 400–1000 nm and ImSpector N17’s
spectral range of about 900–1750 nm for SpectroSort’s VIS and NIR units,
respectively. See Pet. 43. In particular, Petitioner submits that “Carlsohn
calls for sensor units using ImSpector models . . . without specifying which
to use.” Id. (citing in part Ex. 1002, 43). Petitioner further submits that
known options of ImSpector models include ImSpector V10 and N17. Id.
Petitioner asserts that the V10’s spectral range was approximately 400–1000

10 We understand that the claim’s recitation of “spacial” data to be an
inadvertent and harmless typographical error, and the intended term is
properly spelled as “spatial” data. See Pet. 42 (acknowledging the same by
using [sic] in reciting the claim).
IPR2020-00899
Patent 9,950,346 B2

45

nm and the N17’s spectral range was approximately 900–1750 nm. Id.
Petitioner explains that “[b]ecause ImSpector’s V10 spectral range includes
[infrared (“IR”)] and [visible (“VIS”)] wavelengths, the sensor unit using it
meets any single-camera-IR-and-VIS requirement.” Id. (citing Ex. 1008
¶¶ 66–67) (emphasis added). Petitioner reasons that a “POSITA would have
selected ImSpector V10 and N17 for SpectroSort’s VIS and NIR units,
respectively, because they meet SpectroSort’s wavelength specifications and
cover a continuous range from about 400-1750 nm.” Id. (citing Ex. 1002,
41–44, Fig. 17.11; Ex. 1008 ¶¶ 66–67). Petitioner elaborates that a POSITA
“would have understood that doing so ‘would have facilitated accurate
material identification’ by ‘eliminating a gap in coverage between VIS and
NIR wavelengths.’” Id. at 44 (quoting Ex. 1008 ¶ 67).
Patent Owner presents numerous arguments nested in its general
contention that disputes Petitioner’s reason for using the ImSpector V10
sensor in Carlsohn’s system. See PO Resp. 35–38. As explained below,
none of these arguments are persuasive.
Patent Owner submits that “the ImSpector V10 cannot itself be
considered the required ‘single hyperspectral camera.’” Id. at 37. Patent
Owner further submits that Petitioner failed to “prove[] by a preponderance
of the evidence that a POSITA could and would have used the ImSpector
V10 (whatever that is) with the camera chip and lens disclosed in Carlsohn.”
See id.
Patent Owner’s argument is not persuasive. Patent Owner’s argument
relies on physically incorporating ImSpector’s V10 sensor with Carlsohn’s
identification system. Bodily incorporation, however, is not the standard for
obviousness analysis. See In re Mouttet, 686 F.3d 1322, 1332–33 (Fed. Cir.
IPR2020-00899
Patent 9,950,346 B2

46

2012); see also MCM Portfolio LLC v. Hewlett-Packard Co., 812 F.3d 1284,
1294 (Fed. Cir. 2015) (“[t]he test for obviousness is not whether the features
of a secondary reference may be bodily incorporated into the structure of the
primary reference” (citation omitted)), cert. denied, 137 S. Ct. 292 (2016).
In the present case, a POSITA would have been able to use ImSpector V10
and N17 sensors as SpectroSort’s VIS and NIR units, respectively. A
POSITA is “a person of ordinary creativity, not an automaton.” KSR, 550
U.S. at 420–21. We are not persuaded that using the ImSpector V10 and
N17 sensors as SpectroSort’s VIS and NIR units, respectively would have
been “uniquely challenging or difficult for one of ordinary skill in the
art.” Leapfrog Enter. Inc. v. Fisher-Price, Inc., 485 F.3d 1157, 1162 (Fed.
Cir. 2007) (citing KSR, 550 U.S. at 420). Indeed, Carlsohn discloses the
general use of ImSpector sensors, without specifically identifying particular
models. See, e.g., Ex. 1002, 43 (“Imaging spectrographs are commercially
available for visible (VIS) and NIR wavelengths (ImSpector from
Specim)”).
We credit Dr. Arce’s testimony that Carlsohn’s SpectroSort’s material
identification is based on both visible and infrared spectra, and its system
includes both VIS and NIR sensor units. Ex. 1008 ¶ 66 (citing Ex. 1002, 41,
43). Carlsohn is silent as to the particular ImSpector models used, but the
ImSpector V10 and N17 models were available before the critical date of the
’346 patent, and had spectral ranges of 400–1000 nm and 900–1750 nm,
respectively. See id.; see also id. n.3 (explaining that the ImSpector V10
was available before the critical date of the ’346 patent). We agree with and
credit Dr. Arce’s testimony that the “ImSpector V10 and N17 would have
been obvious options for SpectroSort’s VIS and NIR sensor units,” as they
IPR2020-00899
Patent 9,950,346 B2

47

meet Carlsohn’s need for capturing spectra over the required spectral range.
See id. ¶ 66 (citing Ex. 1002, 41–44, Fig. 17.11).
Patent Owner further argues that in addition to Petitioner’s failure to
provide sufficient evidence, “there are two additional reasons the ImSpector
V10, in combination with the asserted system of Carlsohn, fails.” PO Resp.
37. “First, there is no evidence in the record why a POSITA would have
bothered to swap out the VIS and NIR hyperspectral cameras used for the
devices selected by Dr. Arce.” Id. Patent Owner also argues, “Further, if
the two cameras of Carlsohn were replaced by a single camera using a
ImSpector V10, the system would not work as intended . . . [because] Dr.
Arce makes very clear that to achieve sorting . . . a full range of spectra is
needed to be interrogatable, e.g., wavelengths from 400 to 1750 nm.” Id.
Patent Owner’s arguments are not persuasive.
As to Patent Owner’s first argument, that “there is no evidence in the
record why a POSITA would have bothered to swap out the VIS and NIR
hyperspectral cameras” (id.), we disagree. Petitioner cites substantial
evidence.
Petitioner reasons that a “POSITA would have selected ImSpector
V10 and N17 for SpectroSort’s VIS and NIR units, respectively, because
they meet SpectroSort’s wavelength specifications and cover a continuous
range from about 400-1750 nm.” Pet. 43 (citing Ex. 1008 ¶¶ 66, 67; Ex.
1002, 41–44, Fig. 17.11). Petitioner further explains that a “POSITA would
have understood that doing so ‘would have facilitated accurate material
identification’ by ‘eliminating a gap in coverage between VIS and NIR
wavelengths.’” Id. at 44 (quoting Ex. 1008 ¶ 67). Dr. Arce testifies that
“using such sensors would have facilitated accurate material identification.”
IPR2020-00899
Patent 9,950,346 B2

48

Ex. 1008 ¶ 67. Dr. Arce explains that accurate material identification is
facilitated as illustrated by Cowling1 and Arleth, for example.11 See id.
Petitioner cites to Cowling1’s teaching that “VIS through SWIR ‘is
more useful . . . for discriminating materials.’” Pet. 44 (citing Ex. 1018
¶¶ 25–30) (alteration in original). Cowling1 discloses,
The operational waveband of a hyperspectral sensor can
be from the visible (VIS) through to the long-wave infra-red
(LWIR). Experimental measurements indicate that the
visible/short-wave infra-red (VIS/SWIR) region is more useful
than the medium-wave infra-redlong-wave infra-red
(MWIR/LWIR) region for discriminating individual materials
and for Sorting coloured glass. Tests also suggest that the
MWIR/LWIR region is more suited for discriminating between
polymer-coated and non-coated glasses and provides more
separability between other material and plastic and glass classes.
Ex. 1018 ¶ 25 (emphasis added). We agree with Petitioner and credit Dr.
Arce’s testimony that Cowling1 teaches that the visible and short-wave
infra-red region is more useful than other regions for “discriminating
individual materials and for sorting coloured glass.” Id.; see also Pet. 44
(quoting the same); see also Ex. 1008 ¶ 67 (testifying to the same).
Cowling1 is evidence to support Petitioner’s reasoning to use the ImSpector
V10 and N17 sensors.
Petitioner further cites to Arleth’s teaching that “‘wavelengths
rang[ing] from . . . NIR . . . to visible light’ are useful for plastic sorting.”
Pet. 44 (quoting Ex. 1004 ¶ 47). Arleth discloses
To perform the comparison resulting in the determination
of each plastic type, “optically identifiable criteria”, namely rays
of different wavelengths could be evaluated using “sorting
logic”. These wavelengths range from the so-called near infrared

11 US 2007/0029232 A1, published Feb. 8, 2007. Ex. 1018 (“Cowling1”).
IPR2020-00899
Patent 9,950,346 B2

49

(NIR)—with wavelengths reaching from between 0.7 ηm to 2.5
ηm—to visible light with significantly shorter wavelengths.
Ex. 1004 ¶ 47. We agree with Petitioner and credit Dr. Arce’s testimony
and find this paragraph teaches using wavelengths ranging from the near
infrared region to the visible region for determining plastic type. See Pet. 44
(finding the same); see also Ex. 1008 ¶ 67 (testifying to the same). Arleth is
further evidence to support Petitioner’s reasoning to use the ImSpector V10
and N17 sensors.
As to Patent Owner’s second argument, that “if the two cameras of
Carlsohn were replaced by a single camera using an ImSpector V10, the
system would not work as intended. . . [because] Dr. Arce makes very clear
that to achieve sorting . . . a full range of spectra is needed to be
interrogatable, e.g., wavelengths from 400 to 1750 nm” (PO Resp. 37), we
disagree.
Patent Owner’s argument presumes that Carlsohn’s system will not
work if the ImSpector V10 is used, because a “full range of spectra is
needed” with “wavelengths from 400 to 1750 nm.” See id. As explained
above, however, Petitioner’s proposed combination would detect spectra
within this range. We agree with Petitioner’s contention that a “POSITA
would have selected ImSpector V10 and N17 sensors for SpectroSort’s VIS
and NIR units, respectively, because they meet SpectroSort’s wavelength
specifications and cover a continuous range from about 400-1750 nm.” Pet.
43 (citing Ex. 1002, 41–44, Fig. 17.11; Ex. 1008 ¶¶ 66–67). Although
Petitioner proposes to use both the ImSpector V10 and N17 sensors,
Petitioner relies on ImSpector V10’s sensor as “meet[ing the] single-camera-
IR-and-VIS requirement,” “[b]ecause ImSpector V10’s spectral range
includes IR and VIS wavelengths” within its range of 400–1000 nm. See id.
IPR2020-00899
Patent 9,950,346 B2

50

(citing Ex. 1008 ¶¶ 66, 67). For the reasons explained above, we agree with
and credit Dr. Arce’s testimony as to the same. Ex. 1008 ¶¶ 66, 67; see also
id. ¶ 66 (“The ImSpector V10 and N17 would have been obvious options for
SpectroSort’s VIS and NIR sensor units.”).
Patent Owner further argues that the claims of the ’346 patent “require
the ‘single hyperspectral camera’ be able to capture spectra up to at least
1700 nm in wavelength.” PO Resp. 38. For the reasons discussed above in
our Claim Construction, Patent Owner’s argument is premised on a claim
construction that we do not adopt. See supra § II.C.1. Specifically, the
claim does not require the single camera to capture wavelengths up to 1700
nm.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitation 1(b) for the reasons stated by Petitioner
in the Petition, which we adopt as our own findings, as further supported by
the testimony of Dr. Arce.

d) Limitation 1(c) - said camera disposed proximate to a
visible or infrared light source
As discussed above, we construe “disposed proximate to” to mean
“disposed nearby.” See supra § II.D.2. We further note that during
prosecution of the ’346 patent, the applicant acknowledged that the term
“‘proximate’ is broad in its scope.” See Ex. 1020, 70.
Petitioner asserts that SpectroSort’s “VIS and NIR sensor units”
(shown in green in annotated Figure 17.10) “are disposed proximate to a
visible or infrared light source” (shown in purple). Pet. 46; see also supra
(depicting Carlsohn’s annotated Figure 17.10). Petitioner explains that
IPR2020-00899
Patent 9,950,346 B2

51

SpectroSort’s “sensor units are ‘proximate’ to the illumination unit because
they are positioned (1) close thereto (e.g., relative to the length of the
conveyor) and (2) to allow SpectroSort to achieve the ’346 Patent’s
objective: to receive reflected light with ‘sufficient intensity’ to achieve
accurate ‘classification results.’” Pet. 46 (citing Ex. 1002, 43, Figs. 17.9,
17.10).
We reproduce Carlsohn’s Figure 17.9 (Ex. 1002, 42), below:

Figure 17.9 depicts “[s]etup (left) and photograph (right) of an industrial
spectral imaging system (CTR SpectroSort®, 1.8 m observation in length).”
Id. Both images show the camera (“Imaging Spectrograph” or “Sensor
Systems”) and the light source (“Illumination”) of the system.
Patent Owner argues that “[t]he location of the light source, i.e.,
‘illumination,’ which ‘are specially modified floodlights,’ is not specifically
described in Carlsohn.” PO Resp. 39 (citing Ex. 1002, 43). As to
Carlsohn’s Figure 17.9, Patent Owner submits that “the positions of the
illumination components shown in the drawing on the left of Fig. 17.9 may
IPR2020-00899
Patent 9,950,346 B2

52

not be accurately depicted.” Id. at 40 (citing Ex. 2001 ¶ 61) (emphasis
added). In its Sur-Reply, Patent Owner further argues that “Carlsohn shows
a camera above the camera enclosure and thus not ‘proximate’ or ‘nearby’ to
the light source.” PO Sur-Reply 11.
We do not agree with Patent Owner’s arguments.
During prosecution of the ’346 patent, the applicant acknowledged
that the term “‘proximate’ is broad in its scope” (see Ex. 1020, 70), and we
agree with Petitioner that Carlsohn’s light source is also proximate, or
nearby, its camera, as shown in Figure 17.9 (see Pet. 46). Even if Carlsohn’s
camera is above its light source, this arrangement does not preclude
Carlsohn’s camera from also being “proximate” or nearby the light source,
contrary to Patent Owner’s argument. See PO Sur-Reply 11.
Petitioner correctly points out that “[t]he light sources shown in the
’346 patent are in the same relative position” as Carlsohn. See Pet. Reply
12. We further agree with Petitioner that there is nothing in the ’346 patent
that leads us to construe “proximate” as requiring a closer relation between
the camera and light source than what is shown in Carlsohn’s Figure 17.9.
See id.; Ex. 1001, Fig. 1. We credit Dr. Arce’s testimony that a “POSITA
would have understood that SpectroSort’s light sources are ‘proximate’ to
the sensor units, as that term is used in the ’346 Patent, in part because they
are relatively close thereto compared to the length of SpectroSort’s
conveyor,” as shown in Carlsohn’s Figures 17.9 and 17.10. Ex. 1008 ¶ 68.
As correctly pointed out by Dr. Arce, during prosecution of the ’346
patent, the applicant stated that the illumination source is “position[ed] . . .
proximate to the camera so as to achieve ‘high degrees of reflectance.’” Ex.
1020, 70; see also Ex. 1008 ¶ 68 (quoting the same). Carlsohn similarly
IPR2020-00899
Patent 9,950,346 B2

53

discloses that “[t]he light source must provide a spatially sufficiently
homogeneous illumination of the sample with sufficient intensity over the
required spectral range” (Ex. 1002, 43) and we agree with Dr. Arce that in
order for SpectroSort’s light sources to provide sufficient illumination, the
camera must be “proximate” or nearby the illumination (see Ex. 1008 ¶ 68),
just as described by the applicant during prosecution of the ’346 patent (see
Ex. 1020, 70).
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitation 1(c) for the reasons stated by Petitioner
in the Petition, which we adopt as our own findings, as further supported by
the testimony of Dr. Arce.

e) Limitations 1(d) and 1(e) - said light source and said
at least one hyperspectral camera are disposed on the
same side of a stream of said heterogeneous material,
said solid belt having first and second ends, said belt
running beneath said hyperspectral identification
system and upon which said heterogeneous material
is conveyed from said first end to said second end
Petitioner submits that Carlsohn’s SpectroSort is configured to
transport the heterogeneous material from the first end of the conveyor to a
second end, such that there is a stream of material. See Pet. 47 (citing in part
Ex. 1002, 41, 43, Fig. 17.10). Petitioner further submits that “SpectroSort’s
light source and cameras are disposed on the same side of the stream.” Id.
IPR2020-00899
Patent 9,950,346 B2

54

We again reproduce Petitioner’s annotated Figure 17.10 (Pet. 34), below:

Figure 17.10 depicts SpectroSort’s hardware setup, including a conveyor
belt (yellow) transporting material (fraction 1 black pieces, and fraction 2
gray pieces) from the left to the right. See Ex. 1002, 43. The figure depicts
material traveling under an “NIR/VIS sensor system,” a “dark room,” and
“illumination.” Id.
Dr. Arce testifies, “The emitted light reflects off of the waste being
sorted, enters SpectroSort’s dark room, and travels to the [spectral imaging
(“SI”)] system.” Ex. 1008 ¶ 63. Dr. Arce explains, “The measured light
originates from SpectroSort’s illumination unit [purple], which includes an
IPR2020-00899
Patent 9,950,346 B2

55

array of floodlights on each side of SpectroSort’s ‘dark room’ such that the
floodlights and cameras are on the same side of the waste stream.” Id.
(emphasis added). Dr. Arce further testifies that SpectroSort’s SI system
satisfies the claimed “hyperspectral identification system.” See id. ¶¶ 61, 69;
see also Pet. 47.
Patent Owner does not dispute Petitioner’s assertions. See generally
PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitations 1(d) and 1(e) for the reasons stated by
Petitioner in the Petition, which we adopt as our own findings, as further
supported by the testimony of Dr. Arce.

f) Limitation 1(f) - said belt traveling at a preset rate,
said hyperspectral identification system being
disposed over said second end of said belt
Petitioner submits that the SpectroSort’s belt travels at a “preset rate.”
Pet. 47 (citing Ex. 1002, 43). Carlsohn discloses that “SpectroSort is the
transport system providing a relative movement of the samples
perpendicular to the first spatial axis (sampling line) at a constant speed.”
Ex. 1002, 43 (emphasis added).
We agree with Petitioner and find that SpectroSort’s conveyor belt’s
“constant speed” satisfies the claimed “said belt traveling at a preset rate.”
Petitioner further submits that “SpectroSort’s identification system is
over the second end of the belt.” Pet. 47–48 (citing Ex. 1002, Fig. 17.10;
Ex. 1008 ¶¶ 61, 69). Referencing annotated Figure 17.10 (reproduced
supra), Dr. Arce testifies that “SpectroSort includes an identification system
IPR2020-00899
Patent 9,950,346 B2

56

[red outline] that is over the second end of the conveyor belt and has a
spectral imaging (‘SI’) system [green].” Ex. 1008 ¶ 61 (citing Ex. 1002, 41–
43, Fig. 17.10) (alterations in original).
Patent Owner disputes Petitioner’s assertion, arguing that “Carlsohn
does not show or suggest ‘said hyperspectral identification system being
disposed over said second end of said belt.” PO Resp. 43; see also PO Sur-
Reply 11 (“Carlsohn, specifically [Figures] 17.9 and 17.10, do not show the
SpectroSort identification unit as being disposed over the second end of the
belt.” (emphasis omitted)). In support of this position, Patent Owner cites to
the cross examination testimony of Dr. Arce and submits that Dr. Arce
“admits that Carlsohn does not show or suggest ‘said hyperspectral
identification system being disposed over said second end of said belt.’
Rather, he admits that the ‘SpectroSort identification unit is disposed closer
to the terminus of the belt than to the beginning of the belt.[’]” PO Resp. 47
(emphasis added); see also id. at 46–47 (citing Ex. 2023, 108:23–110:24).
Patent Owner further argues that “Petitioner admits that Carlsohn does not
show the hyperspectral identification system disposed over the second end
of the conveyor belt” by presenting its alternative challenge to the claims,
relying on obviousness for this claim limitation rather than disclosure in
Carlsohn in alternative Ground 3. See id. at 49 (citing Pet. 74–80) (emphasis
added).
We disagree with Patent Owner that both Dr. Arce and Petitioner
separately “admit” that Carlsohn does not disclose the limitation.
As to Dr. Arce’s cross-examination testimony, although Dr. Arce
testified that the “end of the belt” is “where it sort of curves” (see Ex. 2024,
110:18–20), he also expressed confusion with Patent Owner’s same line of
IPR2020-00899
Patent 9,950,346 B2

57

questioning (see id. at 110:15–16 (“A. I mean, I’m not sure if I understand
your question.”). With this confusion, we do not find Dr. Arce as admitting
to Patent Owner’s position that Carlsohn’s hyperspectral identification
system is not over the “second end” of its conveyor belt, as that term in used
in claim 1. For example, in his initial declaration, Dr. Arce testifies that
“SpectroSort includes an identification system [red outline] that is over the
second end of the conveyor belt,” referencing the annotated version of
Carlsohn’s Figure 17.10, reproduced above. Ex. 1008 ¶ 61 (alteration in
original).
As to Petitioner’s “admission” that the camera is not over the second
end of the belt, we disagree with Patent Owner’s position that simply
because Petitioner presents an alternative challenge to address the claimed
limitation that Petitioner admits that the first position is flawed. See PO
Resp. 49. Indeed, Petitioner unequivocally submits that “SpectroSort’s
identification system is over the second end of the belt.” Pet. 47–48 (citing
Ex. 1002, Fig. 17.10). Petitioner presents the alternative challenge, in which
Petitioner proposes to move Carlsohn’s identification system closer to the
terminus of the belt, in the event that the “SpectroSort’s identification
system is not sufficiently close to the conveyor’s terminus for (1) it to be
over the conveyor’s ‘second end’ and/or (2) the cameras and light source
that are part of/connected to the identification system to be on the ‘same
side’ of the waste stream.” See id. at 74–75.
Having established that neither Dr. Arce nor Petitioner admits that
Carlsohn’s identification system is not over the second end of the conveyor
belt, we understand that Patent Owner’s argument is one of claim
construction; that is, whether the “second end” of the belt is “immediately
IPR2020-00899
Patent 9,950,346 B2

58

adjacent the ejection system,” as Patent Owner argues. See PO Resp. 13–22.
As explained in our Claim Construction section, we do not construe “said
hyperspectral identification system being disposed over said second end of
said belt” to mean “the hyperspectral identification system is positioned over
the portion of the belt immediately adjacent the ejection system.” See supra
§ II.C.3. Rather, we agree with Petitioner that the “second end” is a term of
approximation, conveying a general location, rather than the terminus of the
belt. See Pet. Reply 15. Indeed, Mr. Smith acknowledged as much during
cross examination. See Ex. 1038, 60:17–21 (“Q When we talk about the
camera being over the second end of the belt, we have to assume, as a
POSITA, that it’s not at the absolute edge of the second belt, don’t we? A I
totally agree.”).
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitation 1(f) for the reasons stated by Petitioner
in the Petition, which we adopt as our own findings, as further supported by
the testimony of Dr. Arce.

g) Limitation 1(g) - a computer configured to receive
and analyze data from said hyperspectral
identification system
Petitioner submits that Carlsohn’s Figure 17.10 depicts a “control
rack” (shown in blue), which is a computer that is configured to receive and
analyze data from its hyperspectral identification system. See Pet. 48; see
also supra (annotated Figure 17.10).
IPR2020-00899
Patent 9,950,346 B2

59

Carlsohn discloses that its control rack contains “a standard state-of-
the-art industry PC . . . for data acquisition and real-time evaluation.” Ex.
1002, 43.
Patent Owner does not dispute Petitioner’s assertions. See generally
PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitation 1(g) for the reasons stated by Petitioner
in the Petition, which we adopt as our own findings, as further supported by
the testimony of Dr. Arce.

h) Limitation 1(h) - to identify target materials of a first
user defined category from among said heterogeneous
materials
Petitioner submits that Carlsohn’s SpectroSort computer is configured
to identify target materials, such as whether the materials are of a particular
type of polymer or whether the paper is “de-inking” or “non-de-inking.”
Pet. 48 (citing in part Ex. 1002, 41, 44–46, Fig. 17.12).
Carlsohn discloses, “The heart of SpectroSort® is a NIR spectral
imaging system that is able to distinguish de-inking from non-de-inking
material based on those differences.” Ex. 1002, 41.
Patent Owner does not dispute Petitioner’s assertions. See generally
PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitation 1(h) for the reasons stated by Petitioner
IPR2020-00899
Patent 9,950,346 B2

60

in the Petition, which we adopt as our own findings, as further supported by
the testimony of Dr. Arce.

i) Limitation 1(i) - identify the spatial position of said
target materials
Petitioner submits that “SpectroSort’s computer is configured to
identify the spatial position of the identified materials based on the position
of the pixels representing them in the hypercube.” Pet. 49 (citing in part Ex.
1002, 44).
Carlsohn discloses, “Spatially resolved material identification and the
subsequent sorting is currently the prevalent SpectroSort application.” Ex.
1002, 44 (emphasis added).
Patent Owner does not dispute Petitioner’s assertions. See generally
PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitation 1(i) for the reasons stated by Petitioner
in the Petition, which we adopt as our own findings, as further supported by
the testimony of Dr. Arce.

j) Limitation 1(j) - to trigger an ejection system at a
preset time delay equal to the distance between said
camera and said ejection system divided by the rate of
travel of the belt
To satisfy this limitation, Petitioner relies on Arleth’s teaching of a
conveyor-speed-based ejection system and reasons that a skilled artisan
would have modified Carlsohn’s SpectroSort system to incorporate a similar
ejection system. See Pet. 37–40, 49–50.
IPR2020-00899
Patent 9,950,346 B2

61

Petitioner submits an annotated version of Arleth’s Figure 4 (id. at
38), which we reproduce, below:

Arleth’s Figure 4 depicts a “schematic representation of a sorting system
with two subsequently arranged sorting devices with the same features.” Ex.
1004 ¶ 27. Plastic parts 315’ are transported on conveyor 521, 531 (shown
in yellow) and sensors 522, 532 (shown in green) identify reflected signals
from the parts, compare them to stored data about known plastics, and
determine if any of the parts are “destined to be ejected.” Id. ¶ 46. Any
parts that are “destined to be ejected” are ejected using nozzles 523, 533
(shown in brown) and a processor “calculates the time and duration of the
control of the ejection nozzle(s) on the basis of the movement parameter of
the conveyor belt.” Id. ¶¶ 46, 49.
Petitioner reasons that “it would have been obvious to configure
SpectroSort’s computer to delay actuation of the ejection system for a time
IPR2020-00899
Patent 9,950,346 B2

62

delay equal to the distance between its cameras . . . and the ejection system
(measured along the direction of belt travel).” Pet. 50. In support of
Petitioner’s reasoning, Dr. Arce testifies that “an appropriate conveyor-
speed-based time delay would have been the distance an object had to travel
from its scanning location divided by the conveyor’s speed.” Ex. 1008 ¶ 93.
Petitioner further reasons that using such conveyor-speed-based time delay
would have permitted proper ejection unit timing without the need for
mechanisms, such as sensors, to detect when scanned material reaches the
ejection unit. Pet. 40. Dr. Arce further testifies that the modification would
“ensure proper ejection timing, especially because SpectroSort’s conveyor
runs ‘at a constant speed.’” Ex. 1008 ¶ 93.
Patent Owner does not dispute Petitioner’s assertions or reasoning for
modifying Carlsohn. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth and as modified satisfies Limitation 1(j) for the reasons
stated by Petitioner in the Petition, which we adopt as our own findings, as
further supported by the testimony of Dr. Arce.

k) Limitation 1(k) - said ejection system disposed
immediately after said second end of said belt,
whereby said desired materials are separated from
said heterogeneous material
Petitioner submits that Carlsohn’s SpectroSort system discloses this
limitation. Pet. 51 (citing in part Ex. 1002, 43).
Carlsohn discloses that its “air-jet array” (shown in brown in
annotated Figure 17.10) “provides the actual separation of material into two
IPR2020-00899
Patent 9,950,346 B2

63

material fractions.” Ex. 1002, 43; supra. In particular, Carlsohn’s Figure
17.10 explicitly shows the air-jet array as separating “fraction 1” onto a first
transport belt and “fraction 2” onto a second transport belt, with the two
transport belts separated by a “separation edge.” See Ex. 1002, 43.
Patent Owner does not dispute Petitioner’s assertions. See generally
PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitation 1(k) for the reasons stated by Petitioner
in the Petition, which we adopt as our own findings, as further supported by
the testimony of Dr. Arce.

l) Limitation 1(l) - said ejection system being triggered
by said computer
Petitioner submits that Carlsohn’s SpectroSort system discloses this
limitation. See Pet. 51.
Carlsohn discloses an “air-jet array” which receives input/output
signals from its computer. See, e.g., Ex. 1002, 46 (“[I]nformation like
position, size, and orientation of an object” obtained by the computer “can
serve as control parameters for an automatic sorter, in this case, an air-jet
array.”).
Patent Owner does not dispute Petitioner’s assertions. See generally
PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth satisfies Limitation 1(l) for the reasons stated by Petitioner
IPR2020-00899
Patent 9,950,346 B2

64

in the Petition, which we adopt as our own findings, as further supported by
the testimony of Dr. Arce.

m) Summary of Independent Claim 1
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by the teachings of Arleth satisfies the limitations of
claim 1 for the reasons stated by Petitioner in the Petition, which we adopt as
our own findings, as further supported by the testimony of Dr. Arce.

4. Dependent Claim 2
Claim 2 depends from claim 1 and further comprises, “a conveyer to
deliver said heterogeneous material to said system.” Ex. 1001, 6:33–35.
To address this claim, Petitioner submits that “SpectroSort comprises
a conveyor that drives its belt to deliver material to the identification
system.” Pet. 51 (citing Ex. 1002, 43; Ex. 1008 ¶ 71).
Carlsohn’s Figure 17.10 depicts a conveyor belt for delivering
material within the system.
As to claim 2, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s teachings satisfies the limitations of claim
2 for the reasons stated by Petitioner in the Petition, which we adopt as our
own findings, as further supported by the testimony of Dr. Arce.

IPR2020-00899
Patent 9,950,346 B2

65

5. Dependent Claim 3
Claim 3 depends from claim 1 and further comprises, “a receiving
system, whereby said materials of a first user defined category are
segregated and collected.” Ex. 1001, 6:36–38.
To address this claim, Petitioner references Carlsohn’s annotated
Figure 17.10 (reproduced supra) and submits that “Carlsohn discloses that
SpectroSort includes a receiving system [pink] comprising two transport
belts and a separation edge.” Pet. 52 (citing Ex. 1002, 43, Fig. 17.10; Ex.
1008 ¶ 71).
Carlsohn’s system receives, segregates, and collects materials into at
least two material fractions; “fraction 1” and “fraction 2.” See Ex. 1002, 43,
Fig. 17.10.
As to claim 3, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s teachings satisfies the limitations of claim
3 for the reasons stated by Petitioner in the Petition, which we adopt as our
own findings, as further supported by the testimony of Dr. Arce.

6. Dependent Claim 4
Claim 4 depends from claim 1 and further comprises, “an array of
lights disposed proximate to said hyperspectral identification system.” Ex.
1001, 6:39–41.
IPR2020-00899
Patent 9,950,346 B2

66

To address this claim, Petitioner submits that “Carlsohn discloses that
SpectroSort’s illumination unit comprises an array of lights.” Pet. 52 (citing
Ex. 1002, 43, Fig. 17.0; Ex. 1008 ¶ 63).
Carlsohn’s Figures 17.9 and 17.10 each depicts at least two sets of
illuminations proximate to, or nearby, its hyperspectral identification system.
As to claim 4, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s teachings satisfies the limitations of claim
4 for the reasons stated by Petitioner in the Petition, which we adopt as our
own findings, as further supported by the testimony of Dr. Arce.

7. Dependent Claim 5
Claim 5 depends from claim 4 and further recites, “wherein said lights
have a wavelength suitable for hyperspectral imaging.” Ex. 1001, 6:42–43.
To address this claim, Petitioner submits that “Carlsohn discloses that
SpectroSort’s lights have a wavelength suitable for hyperspectral imaging
(e.g., are visible and infrared light sources).” Pet. 53 (citing Ex. 1002, 43;
Ex. 1008 ¶ 68).
Carlsohn discloses that its “light sources must provide . . .
illumination of the same with sufficient intensity over the required spectral
range.” Ex. 1002, 43.
As to claim 5, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
IPR2020-00899
Patent 9,950,346 B2

67

We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s teachings satisfies the limitations of claim
5 for the reasons stated by Petitioner in the Petition, which we adopt as our
own findings, as further supported by the testimony of Dr. Arce.

8. Dependent Claim 7
Claim 7 depends from claim 1 and further recites, “wherein said
hyperspectral identification system is configured to collect spectral and
spatial information from heterogeneous material entering said system and
store said spectral and spatial information as images.” Ex. 1001, 6:47–51.
To address this claim, Petitioner submits that “SpectroSort’s
hyperspectral identification system can collect spectral and spatial
information from heterogeneous material entering the system . . . [and it] can
store that information as images to form the hypercube using the push-
broom scanning principle.” Pet. 53 (citing Ex. 1002, 41–42, 44–46, Fig.
17.8; Ex. 1008 ¶¶ 61, 69).
Dr. Arce testifies that “[Carlsohn’s] SI system includes at least one
hyperspectral camera configured to capture and (at least temporarily) store
hyperspectral images of waste that include both spatial and spectral data: its
visible (‘VIS’) and near-infrared (‘NIR’) sensor units.” Ex. 1008 ¶ 61
(citing Ex. 1002, 41–43 (“SI system can feature one or several of these units
(e.g., to cover a wider spectral range),” Figs. 17.8 (depicting the system as
generating a “spectral hypercube” including spatial data along the “1st
spatial axis” and the “2nd spatial axis” and spectral data in each pixel), 17.9,
17.10 (depicting a “VIS sensor Unit” and a “NIR Sensor Unit”)).
IPR2020-00899
Patent 9,950,346 B2

68

As to claim 7, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s teachings satisfies the limitations of claim
7 for the reasons stated by Petitioner in the Petition, which we adopt as our
own findings, as further supported by the testimony of Dr. Arce.

9. Dependent Claim 9
Claim 9 depends from claim 1 and further recites, “wherein said
heterogeneous materials comprise waste and recyclable materials.” Ex.
1001, 6:58–59.
To address this claim, Petitioner submits that “Carlsohn discloses that
the heterogeneous materials SpectroSort can sort comprise waste and
recyclable materials.” Pet. 54 (citing Ex. 1002, 39–41).
Carlsohn discloses that “[t]he main focus of SpectroSort is the sorting
of paper and polymers in the industrial environments” and that the paper and
polymers are then “recycled in a lot of countries in the world.” Ex. 1002,
39–41, 44–46; see also Pet. 54 (finding the same).
As to claim 9, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s teachings satisfies the limitations of claim
9 for the reasons stated by Petitioner in the Petition, which we adopt as our
own findings, as further supported by the testimony of Dr. Arce.
IPR2020-00899
Patent 9,950,346 B2

69

10. Dependent Claim 10
Claim 10 depends from claim 1 and further recites, “wherein said
computer can select a plurality of user defined materials.” Ex. 1001, 6:60–
61.
To address this claim, Petitioner submits that “SpectroSort’s computer
can select a plurality of user-defined materials, such as [de]inking material,
non-deinking material, different polymer types, and (if so-modified)
different glass types.” Pet. 55.
Carlsohn discloses “Sorting of Paper and Polymers” (Ex. 1002, 40)
and that data processing may be performed “using a standard PC” (id. at 44).
As to claim 10, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s teachings satisfies the limitations of claim
10 for the reasons stated by Petitioner in the Petition, which we adopt as our
own findings, as further supported by the testimony of Dr. Arce.

11. Dependent Claim 11
Claim 11 depends from claim 1 and further recites, “wherein said
ejection system comprises a mechanical ejector.” Ex. 1001, 6:62–63.
To address this claim, Petitioner submits that “it would have been
obvious to use a mechanical ejector for SpectroSort’s ejection system.”
Pet. 57 (referencing the rejection of claim 1).
IPR2020-00899
Patent 9,950,346 B2

70

As discussed above, we agreed with Petitioner’s reasoning that a
skilled artisan would have modified Carlsohn’s SpectroSort system to
incorporate a mechanical ejector as taught by Arleth. See supra § II.F.3.j;
see also Pet. 37–40, 49–50.
As to claim 11, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s teachings satisfies the limitations of claim
11 for the reasons stated by Petitioner in the Petition, which we adopt as our
own findings, as further supported by the testimony of Dr. Arce.

12. Dependent Claim 12
Claim 12 depends from claim 1 and further recites, “wherein said
ejection system comprises a fluid jet.” Ex. 1001, 6:64–65.
To address this claim, Petitioner submits that “Carlsohn discloses that
SpectroSort’s ejection system is an air-jet array; because air is a fluid, the
air-jet array satisfies the limitation.” Pet. 56.
Carlsohn discloses that “an airjet array [may] provide[] the actual
separation of material into two material fractions.” Ex. 1002, 43.
As to claim 12, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s teachings satisfies the limitations of claim
12 for the reasons stated by Petitioner in the Petition, which we adopt as our
own findings, as further supported by the testimony of Dr. Arce.
IPR2020-00899
Patent 9,950,346 B2

71

13. Summary of Ground 1
Based on the foregoing, we determine that Petitioner has proven by a
preponderance of the evidence that claims 1–5, 7, and 9–12 are unpatentable
in view of Carlsohn as modified by the teachings of Arleth.

G. Ground 2 – Claim 8 Obvious over Carlsohn, Arleth, Ulrichsen,
and Mayhak
Petitioner submits that claim 8 is unpatentable under 35 U.S.C. § 103
in view of Carlsohn, Arleth, Ulrichsen, and Mayhak. Pet. 57.

1. Ulrichsen (Ex. 1006)
Ulrichsen is a U.S. Patent titled “Determination of Characteristics of
Material.” Ex. 1006, code (54). Ulrichsen discloses a “system for
automatically inspecting matter for varying composition . . . through which
one or more streams of matter are advanced and particular materials therein
are detected through their diffusely reflected IR spectra, if any.” Id. at code
(57). To depict an embodiment of Ulrichsen’s system, we reproduce its
Figure 2, below:
IPR2020-00899
Patent 9,950,346 B2

72

Figure 2 depicts a system for sorting waste objects of different
compositions, with detection from above. Ex. 1006, 10:26–30. In
particular, Figure 2 shows detection unit 120 that receives light reflected
from illuminated waste. Id. at 11:50–58, 12:1–22. To sort the waste,
detection unit 120 controls air nozzles 116 (located at the terminus of
conveyor 104) for directing polymer-coated cartons 125 into bin 126. Id. at
13:13–50.

2. Mayhak (Ex. 1007)
Mayhak is a U.S. Patent titled “Debagger Sorting Machine.” Ex.
1007, code (54). Mayhak discloses a “machine for removing yard waste
from bags and for sorting the waste.” Id. at code (57). To illustrate an
embodiment of Mayhak’s machine, we reproduce its Figure 1, below:
IPR2020-00899
Patent 9,950,346 B2

73

Figure 1 depicts a perspective view of Mayhak’s debagging sorting
machine. Ex. 1007, 5:41–42. In particular, Mayhak discloses debagger
sorter machine 10 supported by a steel frame structure. Id. at 6:15–20. The
machine also has hood 44 mounted above the low end of inclined screen 38.
Id. at 7:21–22. Hood 44 contains fan 47 driven by hydraulic motor 30. Id.
at 7:22–23. Fan 47 lifts bags and bag parts from screen 38 and ejects them
to the rear of the machine. Id. at 7:25–27. Material that does not either fall
through the screen or which was not ejected from the screen by the blower
will vibrate off the low end of the screen, where it can be collected
separately. Id. at 7:28–31.

3. Analysis
Claim 8 depends from claim 1 and further recites, “wherein said
ejection system comprises a plurality of independently controlled air nozzles
disposed beneath said material, and a blower hood disposed above said
IPR2020-00899
Patent 9,950,346 B2

74

material, said blower hood providing an air current whereby said selected
material is blown out of a discharge port.” Ex. 1001, 6:52–57.
To address this limitation, Petitioner proposes to modify Carlsohn’s
SpectroSort system based on the teachings from Ulrichsen and Mayhak. See
Pet. 57–61.
Petitioner relies on Ulrichsen’s teaching of a “‘two step’ air ejection”
that targets materials. See id. at 58 (citing Ex. 1006, 13:26–32). Ulrichsen
discloses that once the materials (in this instance, beverage cartons 125) are
lifted by nozzles 116, “they may be hit with a second air impulse, for
example, a transverse air flow . . . to make them land in a bin at the side of
the conveyor belt 104.” Ex. 1006, 13:26–32; see also id. at Fig. 2
(reproduced supra).
As for Mayhak, Petitioner relies on Mayhak’s teaching of a blower
above screen 38 that has hood 44 and fan 47. See Pet. 58. Mayhak teaches
that bags are lifted from the screen to the blower, which “blows the bag and
bag parts off the screen to one side where they are collected and disposed of
separately.” Ex. 1007, 7:21–27; see also id. at Fig. 1 (reproduced supra).
In modifying Carlsohn’s SpectroSort system based on the teachings of
Ulrichsen and Mayhak, Petitioner reasons that a “POSITA would have
included a blower hood disposed over SpectroSort’s air-jet array . . . to
provide a transverse air flow to eject material lifted by the air-jet array out of
a discharge port.” Pet. 59 (citing Ex. 1008 ¶¶ 100–102). Dr. Arce testifies
that the two-step air ejection method for separating waste material is taught
by Ulrichsen as advantageous. See Ex. 1008 ¶ 100.
As to claim 8, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
IPR2020-00899
Patent 9,950,346 B2

75

We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by the teachings of Arleth, Ulrichsen, and Mayhak
satisfies the limitations of claim 8 for the reasons stated by Petitioner in the
Petition, which we adopt as our own findings, as further supported by the
testimony of Dr. Arce.

H. Ground 3 – Claims 1–7 and 9–16 Obvious over Carlsohn,
Arleth, and Sinram
Petitioner challenges claims 1–7 and 9–16 as being unpatentable
under 35 U.S.C. § 103 in view of Carlsohn, Arleth, and Sinram. Pet. 5.

1. Sinram (Ex. 1003)
Sinram is an international patent publication titled “System and
Method for Identifying and Sorting Material.” Ex. 1003, code (54). Sinram
discloses an “automatic sorting system for identifying and sorting non-
[homogeneous] material” using spectral analysis. See id. at code (57).

2. Analysis
Petitioner largely relies on the same assertions and reasoning
discussed above under Ground 1, namely, Carlsohn and Arleth. See Pet. 62–
65. In further combining Carlsohn and Arleth with Sinram, Petitioner
submits that the proposed modification addresses the limitations of claims 6
and 13–16, which are not addressed under Ground 1. See id. at 66–74.
Petitioner also submits, as an alternative to its challenge of claim 1 under
Ground 1, that a POSITA would have further relied on Sinram’s teachings
IPR2020-00899
Patent 9,950,346 B2

76

and additional rationale for addressing Limitations 1[d] and 1[f]. See id. at
75.

3. Independent Claim 1
Petitioner relies mostly on the same findings and reasoning relied
upon under Ground 1. See Pet. 62–65. As discussed above in connection
with Ground 1, Petitioner asserts that Carlsohn’s “SpectroSort’s
identification system is over the second end of the belt.” Id. at 47–48 (citing
Ex. 1002, Fig. 17.10; Ex. 1008 ¶¶ 61, 69); see also id. at 74 (“SpectroSort’s
identification system is disposed over the conveyor’s second end”).
Petitioner relies on Sinram in the alternative and submits,
To the extent PO argues that SpectroSort’s identification system
is not sufficiently close to the conveyor’s terminus for (1) it to be
over the conveyor’s ‟second end” and/or (2) the cameras and
light source that are part of/connected to the identification system
to be on the ‟same side” of the waste stream, it would have been
obvious to position the identification system closer to the
terminus in Grounds 1-3.
Id. at 75 (citing Ex. 1008 ¶¶ 76–81; Ex. 1003, Fig. 5; Ex. 1002, Fig. 17.10).
In support of its alternative position, Petitioner submits another
annotated version of Carlsohn’s Figure 17.10 to illustrate the proposed
modification. Id. at 76. We reproduce that annotated Figure 17.10, below:
IPR2020-00899
Patent 9,950,346 B2

77

Annotated Figure 17.10 depicts Carlson’s “NIR/VIS sensor system(s),”
“illumination,” and “dark room” shifted closer to the terminus of the
“second end” of the “conveyor belt.” In arriving at this proposed
modification, Petitioner relies on Sinram’s teachings, namely, Sinram’s
Figure 5, and submits an annotated version of this figure. See Pet. 62. We
reproduce Petitioner’s annotated version of Sinram’s Figure 5, below:
IPR2020-00899
Patent 9,950,346 B2

78

Sinram’s “Figure 5 is an enlarged view of a portion of . . . a carriage of the
identification unit.” Ex. 1003, 6. Specifically, Figure 5 depicts system 101
comprising sorting unit 123 and second end 103b of conveyor belt 103. Id.
at 11. Petitioner submits that Sinram’s “identification system 109 is
disposed over belt 103’s second end at the terminus thereof.” Pet. 76.
Petitioner submits that Sinram’s arrangement “confirms that placing
SpectroSort’s identification system closer to the conveyor’s terminus as well
would have been an obvious option.” Id. (citing in part Ex. 1008 ¶ 76).
In support of Petitioner’s proposed modification, Dr. Arce testifies
that “a POSITA would have understood the positioning of Sinram’s
identification unit to be relevant to the positioning of SpectroSort’s
identification system in view of the similarities between the devices.”
Ex. 1008 ¶ 77. Dr. Arce further testifies that a
IPR2020-00899
Patent 9,950,346 B2

79

POSITA would have been motivated to move SpectroSort’s
identification system in this manner to facilitate accurate sorting.
It was known that recyclable material that SpectroSort handles
or would have been obvious for SpectroSort to handle—such as
glass or plastic bottles and paper products—can shift on a
conveyor after being identified by an optical sensor.
Id. ¶ 78. Dr. Arce explains that a “POSITA would have understood that
moving SpectroSort’s system closer to the conveyor’s terminus would
facilitate accurate sorting by reducing the distance between the scanning
point and the air-jet array and thereby mitigat[e] the risk of objects shifting
after being scanned[,] because there would be less time for them to shift.”
Id.
Patent Owner argues that “a POSITA would not be motivated by
Sinram to move the Carlsohn hyperspectral identification system over the
second end of the belt.” PO Resp. 51. In support of this argument, Patent
Owner submits that “[t]he system in Carlsohn utilizes an imaging
spectrograph and a NIR camera to obtain the spectral data . . . while Sinram
does not use a hyperspectral camera.” Id. at 52 (citing in part Ex. 2001
¶¶ 51, 53). Patent Owner further explains that “Sinram uses a combination
of 64 lenses, a multiplexer, and a spectrometer, making the devices
described in Sinram and Carlsohn incompatible at a basic level.” Id. (citing
Ex. 2001 ¶ 54).
We do not agree with Patent Owner’s argument. Even if “Carlsohn
utilizes an imaging spectrograph and a NIR camera,” while “Sinram uses a
combination of 64 lenses, a multiplexer, and a spectrometer,” these are
differences without a distinction. In addressing the particular limitation,
which requires the “hyperspectral identification system being disposed over
said second end of said belt,” Petitioner simply relies on Sinram for teaching
IPR2020-00899
Patent 9,950,346 B2

80

an identification system disposed over the terminus of a second end of a
conveyor belt. See Pet. 76 (citing in part Ex. 1003, Fig. 5). Based on
Sinram’s teaching, Petitioner reasons that a skilled artisan would have
“plac[ed] SpectroSort’s identification system closer to the conveyor’s
terminus” and that this “would have been an obvious option.” Id. (citing in
part Ex. 1008 ¶ 77). Petitioner explains that the modification “would have
facilitated accurate sorting.” Id. at 76–77 (citing Ex. 1008 ¶ 78). Even if we
agree with Patent Owner’s extensive summary outlining the differences
between Carlsohn’s system and Sinram’s system (see, e.g., PO Resp. 53–
56), these differences do not persuade us that a skilled artisan would not
have modified Carlsohn based on Sinram’s teachings, and as proposed by
Petitioner. At the end of the day, Sinram’s Figure 5 depicts identification
system 109 over the terminus of the second end of conveyor belt 103 (see
Ex. 1008 ¶ 73 (depicting Sinram’s annotated Fig. 5)), and Patent Owner’s
extensive argument regarding the differences between Carlsohn’s and
Sinram’s particular identification systems are inapposite.
Patent Owner further argues that “process-time concerns in the
Carlsohn system preclude considering Sinram’s identification-system-at-the-
terminus sorter.” PO Resp. 58 (citing Ex. 2022 ¶ 25). Patent Owner
contends that “Carlsohn’s identification system takes at best two minutes to
process the image and identify the selected material” (id.) and that
“Carlsohn’s identification system cannot process the large amounts of data a
hyperspectral system of the ’346 Patent would” (id. at 59). Patent Owner
argues that as a result, “a POSITA would not be motivated in view of
Sinram to move the identification system in Carlsohn to the second end of
the belt.” Id. (citing Ex. 2001 ¶ 65; Ex. 2022 ¶ 22).
IPR2020-00899
Patent 9,950,346 B2

81

We do not agree with Patent Owner’s second argument. Rather, we
agree with Petitioner that a “POSITA would know that moving the imaging
system closer to the terminus of the belt would involve a tradeoff, either
requiring higher processing power to process images more quickly,
processing fewer images, and/or slowing the belt to allow the processing to
take place in order to engage the ejector system.” Pet. Reply 17.
We first point out that Carlsohn discloses that “[f]or the SpectroSort
system, the time for acquisition, processing, data evaluation, and
post[-]processing of one sample line is less than 15 ms.” Ex. 1002, 44
(emphasis added); see also Ex. 1008 ¶ 79 (testifying that Carlsohn’s 15 ms
scan time is about the same as Sinram’s 14.3 ms scan time). 15 ms
processing time is a far cry from Patent Owner’s calculation of two minutes,
and Patent Owner’s attempt to ignore Carlsohn’s 15 ms teaching presumes,
without sufficient basis, that “Carlsohn’s identification system cannot
process the large amounts of data” in that time. See PO Resp. 59.
As explained above, Patent Owner’s argument is premised on
Carlsohn’s system needing “at best two minutes to process the image and
identify the selected material.” PO Resp. 58 (citing Ex. 2001 ¶ 65). We
disagree with Patent Owner’s assertion that at least two minutes is required.
In support of Patent Owner’s two-minute processing time, Mr. Smith
testifies that
Table 17.2 of Carlsohn gives the error rates and processing times
for various experiments classifying tomatoes to train the system.
According to this table, it would take the system on average 430
seconds (7.2 minutes) to classify each tomato if all 186 bands
(i.e., 186 reflection values between 490 nm and 730 nm) are
used. Additionally, the error rate would be about 10% to 12%.
If only 19 discrete bands between 490 nm and 730 nm are used,
then it would take the system 120 seconds (2 minutes) to classify
IPR2020-00899
Patent 9,950,346 B2

82

each tomato, with an error rate of 7% to 10%. If Carlsohn is
moved to the terminus of the belt like Sinram, the belt would
have to stop moving for about 2 minutes after each row of items
is scanned or the belt would have to move extremely slow (way
slower than the desired feed rate for recycling facilities) such that
Carlsohn has enough time to classify the items in the row and
determine which items to keep and which items to eject.
Ex. 2001 ¶ 65 (citing Ex. 1002, 30, 38). As explained above, Mr. Smith’s
testimony is based on Carlsohn’s Figure 17.2. We further reproduce
Carlsohn’s Table 17.2, below:

Although Figure 17.2 depicts “Error Rates for Five Tomatoes for a Varying
Number of Wavelength Bands (Features), Using Parzen Classification,” with
“Processing Times” of between 120 to 430 seconds, this tomato experiment
is not a teaching that Carlsohn’s SpectroSort system needs “at best two
minutes to process the image and identify the selected material,” as Patent
Owner and Mr. Smith contend. See PO Resp. 58 (citing Ex. 2001 ¶ 65).
Upon close examination of this Table, upon which Mr. Smith’s testimony
relies (Ex. 2001 ¶ 65), the “data used in this experiment are a random
selection of 1000 pixels from spectral images of five tomatoes in five
ripeness classes . . . . The processing time is the mean of the elapsed time
needed for training the Parzen classifier per tomato.” Ex. 1002, 38
IPR2020-00899
Patent 9,950,346 B2

83

(emphases added). Even if it may take two minutes to classify each tomato,
with varying ripeness and under this particular experiment, this fact is
inapposite; Petitioner relies on Carlsohn’s SpectroSort system for sorting
and classifying paper, plastic, and glass—not tomatoes. See, e.g., Pet. 44
(“And it would have permitted SpectroSort to sort glass too, e.g., by color
and/or to remove ceramic glass, which a POSITA would have been
motivated to do”); see also Ex. 1008 ¶ 78 (“POSITA would have been
motivated to move SpectroSort’s identification system in this manner to
facilitate accurate sorting. It was known that recyclable material that
SpectroSort handles or would have been obvious for SpectroSort to
handle—such as glass or plastic bottles and paper products—can shift on a
conveyor after being identified by an optical sensor.”).
To reiterate, Carlsohn discloses that “[f]or the SpectroSort system, the
time for acquisition, processing, data evaluation, and post[-]processing of
one sample line is less than 15 ms.” Ex. 1002, 44 (emphasis added).
Patent Owner discounts Carlsohn’s 15 ms teaching and argues that
“[e]ven in view of Sinram, Carlsohn’s hyperspectral identification system
could not be placed over the second end of the belt because around the time
of the Critical Date, November 18, 2007, there were no adequately powerful
computer and algorithm options to analyze desired materials before reaching
the ejection system.” PO Resp. 60 (citing Pet. 78); see also id. at 59
(acknowledging the 15 ms processing time but nevertheless arguing that
“Carlsohn’s identification system cannot process the large amounts of data”
in that time).
We do not agree with Patent Owner’s argument. Again, Carlsohn
teaches that “the SpectroSort system, the time for acquisition, processing,
IPR2020-00899
Patent 9,950,346 B2

84

data evaluation, and post[-]processing of one sample line is less than 15 ms.”
Ex. 1002, 44. We credit Dr. Arce’s testimony and find that a processing
time of less than 15 ms is sufficiently fast to support Petitioner’s proposed
modification, and Patent Owner cites nothing in the record to persuade us
that Carlsohn’s express teaching is wrong. See PO Resp. 59–61. Carlsohn
is clear in its teaching of a “less than 15 ms” processing time for its
SpectroSort system. See Ex. 1002, 44.
Mr. Smith’s opposite conclusion, that “[e]ven today, the hyperspectral
camera technology requires a slower belt speed 2.5 m/s versus 3.0 m/s for
the multiplexed technology,” is conclusory and unsupported by evidence.
Ex. 2022 ¶ 25. Mr. Smith’s declaration incorporates an image of
unspecified origin, which he characterizes as illustrating a “recent version of
an LLA Instruments NIR hyperspectral imaging camera [that] is not over the
end of the conveyor belt.” Id. Even taking this image at face value,
however, one example of one sorting system that has a hyperspectral
imaging camera disposed upstream of the belt terminus does not establish
that Carlsohn was wrong to describe its system as having a processing time
of 15 ms.
Finally, Patent Owner argues that “a POSITA would not be motivated
to consider Sinram, as its commercial embodiment teaches away from the
use of a hyperspectral camera in identifying desired materials in a
heterogeneous stream.” PO Resp. 61 (citing Ex. 2001 ¶¶ 55–57) (emphasis
added).
We do not agree with Patent Owner’s final argument, as Patent Owner
has not identified anything in Sinram (or Carlsohn, for that matter) that
criticizes, discredits, or otherwise discourages moving Carlsohn’s
IPR2020-00899
Patent 9,950,346 B2

85

identification system closer to the terminus of the second end of the
conveyor belt. See In re Fulton, 391 F.3d 1195, 1201 (Fed. Cir. 2004) (“The
prior art’s mere disclosure of more than one alternative does not constitute a
teaching away from any of these alternatives because such disclosure does
not criticize, discredit, or otherwise discourage the solution claimed.”).
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by the teachings of Arleth and Sinram satisfies the
limitations of claim 1 for the reasons stated by Petitioner in the Petition,
which we adopt as our own findings, as further supported by the testimony
of Dr. Arce.

4. Dependent Claims 2–5, 7, and 9–11
In challenging dependent claims 2–5, 7, and 9–11 under Carlsohn,
Arleth, and Sinram, Petitioner relies on the same argument and evidence
discussed above under Ground 1. See Pet. 62–80.
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s and Sinram’s teachings satisfies the
limitations of claims 2–5, 7, and 9–11 for the reasons stated by Petitioner in
the Petition, which we adopt as our own findings, as further supported by the
testimony of Dr. Arce.

IPR2020-00899
Patent 9,950,346 B2

86

5. Dependent Claim 6
Claim 6 depends from claim 4 and further recites, “wherein said
computer comprises software configured to compare said images to images
stored in a library of images.” Ex. 1001, 6:44–46.
To address claim 4, Petitioner submits that the combination of
Carlsohn, Arleth, and Sinram renders claim 4 obvious for the same reasons
that Carlsohn and Arleth renders claim 4 obvious. See Pet. 66.
As to claim 6, Petitioner reasons that “it would have been obvious to
modify SpectroSort to capture and store images in a library as reference data
such that its computer can compare the hyperspectral images captured
during sorting.” Id. (citing Ex. 1008 ¶¶ 82–85). Dr. Arce testifies that
“Sinram confirms that this would have been an obvious option,” citing
Sinram’s “learn mode.” Ex. 1008 ¶ 84. Dr. Arce further testifies that
Obtaining the reference spectral data with the same device used
to identify materials would mitigate the risk of the measured data
and the reference data exhibiting differences attributable to using
different equipment. Implementing such a “learning mode” in
SpectroSort would have been straightforward and routine for a
POSITA, as it would have called for using the same equipment
and storing the obtained information in a library in SpectroSort’s
computer.
Id. ¶ 85.
As to claim 6, Patent Owner does not present additional arguments
beyond those presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s and Sinram’s teachings satisfies the
limitations of claim 6 for the reasons stated by Petitioner in the Petition,
IPR2020-00899
Patent 9,950,346 B2

87

which we adopt as our own findings, as further supported by the testimony
of Dr. Arce.

6. Dependent Claims 12 and 13
Claim 12 depends from claim 1 and further recites, “wherein said
ejection system comprises a fluid jet.” Ex. 1001, 6:64–65. Claim 13
depends from claim 12 and further recites, “wherein said fluid jet is a water
jet.” Id. at 6:66–67.
To address these claims, Petitioner reasons that “it would have been
obvious to use Sinram’s nozzles 131 in SpectroSort’s air jet array” and that
nozzles 131 “can produce a water jet.” See Pet. 73 (citing Ex. 1008 ¶ 87).
Dr. Arce testifies that Sinram’s nozzles 131 are “tapered to produce a
strong air jet.” Ex. 1008 ¶ 87. Dr. Arce further testifies that a “POSITA
would have been motivated to use nozzles 131 because ‘the extremities of
the air jets 131 are [] tapered so as to increase thrust of air coming out of
said air jets,’ which would ensure that SpectroSort’s air-jet array provides
sufficient force to achieve material separation.” Id. ¶ 86 (alteration in
original). Dr. Arce also addresses using water jets as an obvious alternative
to air jets. See id. ¶ 87.
As to claims 12 and 13, Patent Owner does not present additional
arguments beyond those presented to its independent claim 1. See generally
PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s and Sinram’s teachings satisfies the
limitations of claims 12 and 13 for the reasons stated by Petitioner in the
IPR2020-00899
Patent 9,950,346 B2

88

Petition, which we adopt as our own findings, as further supported by the
testimony of Dr. Arce.

7. Independent Claim 14
In challenging independent claim 14, Petitioner relies largely on the
same findings and reasoning relied on in challenging independent claim 1
and dependent claim 6. See Pet. 66–71.

a) Limitation 14(p) - A method for sorting heterogeneous
material
Petitioner submits that Carlsohn’s SpectroSort system is a system for
sorting heterogeneous material, referencing Limitation 1(a). Pet. 66 (citing
Ex. 1002, 39–41, 44–46; Ex. 1008 ¶¶ 57–59, 70–71).
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(p), to the
extent that the preamble is limiting, for the reasons stated by Petitioner in the
Petition, which we adopt as our own findings, as further supported by the
testimony of Dr. Arce.

IPR2020-00899
Patent 9,950,346 B2

89

b) Limitation 14(a) - reflecting visible or infrared light
from the surface of said heterogeneous material
disposed on a conveyer from a light source
Petitioner submits that “[d]uring sorting, SpectroSort’s light source
‘provide[s] . . . illumination of the sample [on the conveyor],” referencing
Limitations 1(a)–1(c). Pet. 66–67 (second to fourth alterations in original).
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(a) for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

c) Limitation 14(b) - [the light source] disposed
proximate to a hyperspectral imager
Petitioner submits that “SpectroSort’s light source is disposed
proximate to a hyperspectral imager: the SI system with its sensor units,”
referencing Limitation 1(c). Pet. 67.
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(b) for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

IPR2020-00899
Patent 9,950,346 B2

90

d) Limitation 14(c) - conveying said heterogeneous
material at a predetermined rate beneath said
hyperspectral imager
Petitioner submits that “SpectroSort’s belt conveys the heterogeneous
material at a predetermined rate beneath the hyperspectral imager,”
referencing Limitation 1(f). Pet. 67.
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(c) for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

e) Limitation 14(d) - generating hyperspectral images
containing both spectral and spatial data of said
heterogeneous material by receiving a plurality of
selected spectral bands of infrared and visible light
corresponding to spectral signatures of target
materials to be identified if present in said
heterogeneous material as said heterogeneous
material passes on said conveyer beneath said
hyperspectral imaginer
Petitioner submits that “[d]uring sorting, SpectroSort generates
hyperspectral images containing spectral and spatial data to form a
hypercube,” referencing Limitation 1(b). Pet. 68.
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
IPR2020-00899
Patent 9,950,346 B2

91

in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(d) for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

f) Limitation 14(e) - comparing said hyperspectral
images of said heterogeneous material to
hyperspectral images of known materials
Petitioner submits that “Carlsohn compares each hypercube pixel’s
spectrum to reference data for material identification,” citing claim 6. Pet.
69.
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(e) for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

g) Limitation 14(f) - identifying said target materials
Petitioner submits that “[d]uring sorting, SpectroSort identifies target
materials (e.g., certain polymer types, ‘de-inking’ or ‘non-de-inking’ paper,
and/or (if so-modified) glass types,” referencing Limitations 1(b) and 1(h).
Pet. 69.
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
IPR2020-00899
Patent 9,950,346 B2

92

in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(f) for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

h) Limitation 14(g) - activating an ejection system with a
computer
Petitioner submits that “[d]uring sorting, SpectroSort’s computer
activates an ejection system, which is an air-jet array,” referencing
Limitations 1(j) and 1(l). Pet. 69.
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(g) for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

i) Limitation 14(h) - said computer being configured to
physically isolate target material with said ejection
system from said heterogeneous material
Petitioner references Limitation 1(k) and submits that
SpectroSort’s computer is configured to physically isolate target
materials, such as de-inking paper or certain polymer types from
the heterogeneous material by controlling the air-jet array such
that certain materials (if not acted on) fall to the “fraction 1”
transport belt and others (if acted on) clear the separation edge to
reach the “fraction 2” transport belt.
Pet. 70 (citations omitted).
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
IPR2020-00899
Patent 9,950,346 B2

93

We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(h) for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

j) Limitation 14(i) - by locating said target material
within said heterogeneous material
Petitioner submits that “[d]uring sorting, SpectroSort’s computer uses
the captured spatial data to determine material positioning on the belt based
on the position of the hypercube’s pixel(s) representing the materials,”
referencing Limitations 1(b) and 1(i). Pet. 70.
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(i) for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

k) Limitation 14(j) - and [by] ejecting said target
material with said ejection system at a preset time
delay equal to a distance between said hyperspectral
imager and said ejection system divided by a rate of
conveying said heterogeneous material
Petitioner reasons that
it would have been obvious to configure SpectroSort’s computer
to delay actuation of the ejection system until a preset time equal
to the distance between its hyperspectral imager (e.g., from the
SI system, the VIS sensor unit, or the NIR sensor unit) and the
IPR2020-00899
Patent 9,950,346 B2

94

ejection system (measured along the direction the belt travels in)
divided by the conveyor’s speed, which meets the claim
limitation.
Pet. 71 (referencing in relevant part Limitation 1(j)).
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that Carlsohn
in view of Arleth’s and Sinram’s teachings satisfies Limitation 14(j) for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

l) Summary of Independent Claim 14
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by the teachings of Arleth and Sinram satisfies the
limitations of claim 14 for the reasons stated by Petitioner in the Petition,
which we adopt as our own findings, as further supported by the testimony
of Dr. Arce.

8. Dependent Claim 15
Claim 15 depends from claim 14 and further recites, “wherein said
hyperspectral images comprise spectral and spatial data for articles within
said heterogeneous material.” Ex. 1001, 8:12–14.
To address this claim, Petitioner submits that “[t]he hyperspectral
images SpectroSort captures include spectral and spatial data for paper,
polymeric, and/or (if so-modified) glass . . . articles.” Pet. 71–72 (citing Ex.
1002, 40–41, 44–46, Fig. 17.12; Ex. 1008 ¶¶ 57–59, 61–64, 69–70).
IPR2020-00899
Patent 9,950,346 B2

95

Carlsohn discloses using both “spectral imaging” and “spatial
information” to sort paper, waste, and other materials. See Ex. 1002, 41.
As to claim 15, Patent Owner does not present additional arguments
beyond those presented to its independent claim 14 (vis-à-vis claim 1). See
generally PO Resp.
We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s and Sinram’s teachings satisfies the
limitations of claim 15 for the reasons stated by Petitioner in the Petition,
which we adopt as our own findings, as further supported by the testimony
of Dr. Arce.

9. Dependent Claim 16
Claim 16 depends from claim 14 and further recites, “wherein said
heterogeneous material comprises waste material and recyclable material in
combination.” Ex. 1001, 8:15–17.
To address this claim, Petitioner submits that “Carlsohn discloses that
the heterogeneous materials SpectroSort can sort comprise waste and
recyclable materials.” Pet. 72 (citing in part Ex. 1002, 39–41).
Carlsohn discloses that “[t]he main focus of SpectroSort is the sorting
of paper and polymers in the industrial environments,” which are then
“recycled in a lot of countries in the world.” Ex. 1002, 39–41, 44–46.
Carlsohn further discloses SpectroSort as a “waste sorting system.” See id.
at 41 (“The core module of CRTs waste sorting system SpectroSort . . . .”).
As to claim 16, Patent Owner does not present additional arguments
beyond those presented to its independent claim 14. See generally PO Resp.
IPR2020-00899
Patent 9,950,346 B2

96

We are persuaded by Petitioner’s argument and evidence, and find
that Petitioner has established by a preponderance of the evidence that
Carlsohn as modified by Arleth’s and Sinram’s teachings satisfies the
limitations of claim 16 for the reasons stated by Petitioner in the Petition,
which we adopt as our own findings, as further supported by the testimony
of Dr. Arce.

10. Summary of Ground 3
Based on the above, we determine Petitioner has proven by a
preponderance of the evidence that claims 1–7 and 9–16 are unpatentable
over Carlsohn in view of the teachings of Arleth and Sinram.

I. Ground 4 – Claim 8 Obvious over Carlsohn, Arleth, Ulrichsen,
Mayhak, and Sinram
Claim 8 depends from claim 1. See Ex. 1001, 6:52. In addressing
claim 8 as obvious over Carlsohn, Arleth, Ulrichsen, Mayhak, and Sinram,
Petitioner relies on the combination of Carlsohn, Arleth, and Sinram for
addressing independent claim 1, under Ground 3, and further relies on
Ulrichsen and Mayhak for addressing the particular limitations of claim 8,
under Ground 2.
Patent Owner does not present additional arguments beyond those
presented to its independent claim 1. See generally PO Resp.
For the reasons discussed above, we are persuaded by Petitioner’s
argument and evidence and find that Petitioner has established by a
preponderance of the evidence that Carlsohn in view of Arleth, Ulrichsen,
Mayhak, and Sinram satisfies claim 8 for the reasons stated by Petitioner in
IPR2020-00899
Patent 9,950,346 B2

97

the Petition, which we adopt as our own findings, as further supported by the
testimony of Dr. Arce.

J. Grounds 5 and 6 – Claim 2 Obvious over Carlsohn, Arleth, and
Kampf (Ground 5) and Carlsohn, Arleth, Sinram, and Kampf (Ground
6)
Ground 5 is based on Carlsohn, Arleth, and Kampf and Ground 6 is
based on Carlsohn, Arleth, Sinram, and Kampf. See Pet. 5–6.

1. Kampf (Ex. 1005)
Kampf is a U.S. Patent Application titled “Method for the Automatic
Analysis of Refuse Containing Polymers and an Automatic Analytical
Device for this Purpose.” Ex. 1005, code (54). Kampf describes a system
and method for analyzing refuse objects containing polymers, such as
carpets, in a recycling plant. See id. at code (57), Fig. 1.

2. Analysis
Claim 2 depends from claim 1 and further comprises “a conveyor to
deliver said heterogeneous material to said system.” Ex. 1001, 6:33–35.
Petitioner presents this alternative challenge to claim 2 “[t]o the
extent PO argues that claim 2 requires a conveyor in addition to one driving
claim 1’s belt, it would have been obvious to include one in Grounds 1 and 3
in view of Kampf.” Pet. 80.
Petitioner relies on Kampf for teaching the additional conveyor, citing
Figure 1 of Kampf, which we reproduce, below:
IPR2020-00899
Patent 9,950,346 B2

98

Figure 1 shows a first embodiment of Kampf’s refuse recycling plant
with an automatic analytical device. Ex. 1005 ¶ 36. Kampf discloses
system 1 including conveyor 4 that receives unsorted material from trucks 3
and delivers it to conveyor 6 for sorting. Id. ¶ 43.
Petitioner reasons that a skilled artisan would have included an
additional conveyor to Carlsohn’s SpectroSort’s system “to rapidly deliver
waste material to the [SpectroSort system’s] conveyor.” Pet. 81 (citing Ex.
1008 ¶ 97).
Patent Owner does not offer any additional arguments specifically
addressing the combination with Kampf. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that:
(1) Carlsohn in view of Arleth and Kampf (Ground 5) and Carlsohn in view
IPR2020-00899
Patent 9,950,346 B2

99

of Arleth, Sinram, and Kampf (Ground 6) satisfy claim 2 for the reasons
stated by Petitioner in the Petition, which we adopt as our own findings, as
further supported by the testimony of Dr. Arce.

K. Grounds 7 and 8 – Claims 12 and 13 over Carlsohn, Arleth,
and Cowling2 (Ground 7) and Carlsohn, Arleth, Sinram, and
Cowling2 (Ground 8)
Ground 7 is based on Carlsohn, Arleth, and Cowling2, and Ground 8
is based on Carlsohn, Arleth, Sinram, and Cowling2. See Pet. 6.

1. Cowling2 (Ex. 1019)
Cowling2 is a U.S. Patent Application titled “Separating Device and
Sorting Apparatus with Two-Dimensionals [sic] Array of Nozzles and
Method of Sorting Objects.” Ex. 1019, code (54). Cowling2 describes a
sorting apparatus for sorting objects in a stream, such as a stream of waste,
according to size, shape, object-material, and color. See id. ¶ 1.

2. Analysis
Claim 12 depends from claim 1 and further recites, “wherein said
ejection system comprises a fluid jet.” Ex. 1001, 6:64–65. Claim 13
depends from claim 12 and further recites, “wherein said fluid jet is a water
jet.” Id. at 6:66–67.
Petitioner submits this alternative challenge to claims 12 and 13 in the
event that Patent Owner argues that Carlsohn’s SpectroSort’s air-jet array
and Sinram’s nozzles do not satisfy the claims. Pet. 82.
IPR2020-00899
Patent 9,950,346 B2

100

Petitioner relies on Cowling2’s teaching of a waste sorting system that
uses air or water jets for separating waste material. Id. (citing in part
Ex. 1019 ¶¶ 1, 2, 6, 7, 35). We reproduce Cowling2’s Figure 1, below:

Figure 1 depicts Cowling2’s sorting apparatus. Ex. 1019 ¶ 29. In
particular, Figure 1 shows apparatus 10 with a conveyor belt and air-
separation units 17, 18, 20, and 21, each having upwardly-directed nozzles.
Id. ¶ 35. Cowling2 explains, “The direction and air speed of the air jets is
sufficient to blast the plastics material object off the upper surface 12 of the
conveyor belt and onto the orthogonal belt 26a and thence into the receptacle
26.” Id. ¶ 37.
Cowling2 further discloses that “water jets may be [utilized] . . .
where the nature of the objects so dictates.” Id. ¶ 7.
Petitioner reasons that a skilled artisan would have modified
SpectroSort’s system by incorporating Cowling2’s water jets, as “water jets
would have imparted larger forces to facilitate material separation,
particularly for ‘larger and/or heavier objects SpectroSort is designed to
handle.’” Pet. 83.
IPR2020-00899
Patent 9,950,346 B2

101

Patent Owner does not offer any additional arguments specifically
addressing the combination with Cowling2. See generally PO Resp.
We are persuaded by Petitioner’s argument and evidence and find that
Petitioner has established by a preponderance of the evidence that: (1)
Carlsohn in view of Arleth and Cowling2 (Ground 7) and Carlsohn in view
of Arleth, Sinram, and Cowling2 (Ground 8) satisfy claims 12 and 13 for the
reasons stated by Petitioner in the Petition, which we adopt as our own
findings, as further supported by the testimony of Dr. Arce.

IPR2020-00899
Patent 9,950,346 B2

102

III. PETITIONER’S MOTION TO EXCLUDE
Petitioner filed a Motion to Exclude Evidence. Paper 34 (“Mot.” or
“Motion”). Patent Owner filed an Opposition to the Motion (Paper 35,
“Opp.” or “Opposition”) and Petitioner filed a Reply to the Opposition
(Paper 37, “Reply Opp.”).

A. Analysis
As the moving party, Petitioner has the burden of proof to establish
that it is entitled to the requested relief. See 37 C.F.R. § 42.20(c).
Petitioner moves to exclude certain parts of Mr. Smith’s declaration
testimony (Ex. 2022, “Declaration” or “Dec.”). Mot. 1. In particular,
Petitioner seeks to exclude paragraphs 6, 12, 15–21, 25 and Exhibits E, F,
and G from the Declaration. Id.

1. Exhibit E, Paragraph 12
Mr. Smith testifies that “[t]here is nothing about [Carlsohn’s title] that
would indicate to a POSITA that this book would contain anything relevant
to hyperspectral imaging.” Ex. 2022 ¶ 12. Mr. Smith further testifies, “I
have attached as Exhibit E a list of articles that are pre-2008 regarding some
aspect of ‘color image processing,’” which is the title of the Carlsohn Book.
Id. Exhibit E is a list of 48 documents with the heading, “Documents on
Color Image Processing that Don’t Have Anything on Hyperspectral
Imaging.” Ex. 2022, 19–21.
Petitioner seeks to exclude “Exhibit E and its accompanying
declaration testimony . . . because they do not tend to show that public
accessibility of Carlsohn (Exhibit 1002) was more or less probable” under
Federal Rules of Evidence (“FRE”) 401. Mot. 4. Petitioner submits that Mr.
IPR2020-00899
Patent 9,950,346 B2

103

Smith’s “opinion does not tend to prove or disprove that Carlsohn would
have been publicly accessible to the public” (id.) and that “[t]here are too
many inferential jumps required here to meet the threshold for relevance”
(id. at 5 (citation omitted)).
Petitioner also seeks to exclude Exhibit E and its accompanying
testimony under FRE 702 as unreliable expert testimony. See Mot. 10–11
(citing in part Daubert v. Merrill Dow Pharm., 509 U.S. 579, 595 (1993)).
Petitioner submits that “Mr. Smith did not provide enough information about
his knowledge or search methodologies to evaluate whether he is qualified to
render an opinion, nor whether the limited queries he employed were
sufficient, or reliable.” Id. at 11.
In addition to FRE 401 and 702, Petitioner seeks to exclude Exhibit E
and its accompanying testimony under FRE 703, explaining that “the
proponent of the opinion may disclose [Exhibit E] to the jury only if [its]
probative value in helping the jury evaluate the opinion substantially
outweighs [its] prejudicial effect.” Mot. 14 (quoting FRE 703).
As to Exhibit E and its accompanying testimony (Ex. 2022 ¶ 12), we
deny Petitioner’s Motion. Petitioner’s arguments go to weight, not
admissibility, and the Board is capable of assessing the weight to be given
the testimony without excluding it.
Regarding relevance, and as correctly argued by Patent Owner,
Exhibit E and paragraph 12 are relevant to refute Dr. Arce’s testimony as to
whether Carlsohn was publicly accessible before the critical date of the
invention. See Opp. 2–3. Indeed, this is a central issue in our proceeding.
See, e.g., PO Resp. 24–35; see also supra § II.D.
IPR2020-00899
Patent 9,950,346 B2

104

As to reliability, the policy considerations for excluding expert
testimony, such as those implemented by the gatekeeping framework
established by the Supreme Court in Daubert, are less compelling in bench
proceedings such as inter partes reviews. Unlike a lay jury, the Board by
statutory definition has competent scientific ability (35 U.S.C. § 6) and has
significant experience in evaluating expert testimony. See Nestle Healthcare
Nutrition, Inc. v. Steuben Foods, Inc., IPR2015-00249, Paper 76 at 23
(PTAB June 2, 2016); see also Tr. 28:14–29:16. Accordingly, the danger of
prejudice in this proceeding is considerably lower than in a conventional
district court trial in front of a lay jury.
Accordingly, we deny Petitioner’s motion to exclude Exhibit E and
paragraph 12 of Mr. Smith’s declaration (Ex. 2022).

2. Exhibits F, G, and Paragraphs 15–21
Exhibit F is “Google Search Pages” and Exhibit G is “Yahoo Search
Pages.” Ex. 2022, 22–202. As to Exhibit F, Mr. Smith testifies that a
“Google search was made to search for web pages dated in the year of 2007
and a separate search made for each of the terms above that Dr. Arce said he
would use to learn about hyperspectral imaging and waste material sorting.”
Ex. 2022 ¶ 15. Mr. Smith testifies that “[t]here was not a single search result
that returned the Carlsohn[] reference” (id. ¶ 16 (emphasis omitted)) and that
this search “would be duplicable by anybody with an internet connection”
(id. ¶ 17). As to Exhibit G, Mr. Smith presents similar testimony regarding
the Yahoo search. See id. ¶¶ 18–19. Mr. Smith further testifies that these
search results “demonstrate[] that reasonable diligence from a POSITA
would not have found the Carlsohn reference prior to the critical date.” Id.
¶ 20.
IPR2020-00899
Patent 9,950,346 B2

105

Petitioner moves to exclude Exhibits F, G and the accompanying
testimony under FRE 401. Mot. 6. Petitioner submits that this evidence
“provides an incomplete view of what a POSITA could access to ascertain
the prior art” (id.) and “the absence of the reference in the limited search
results provided[] does not negate the fact that the publication could be
discoverable by other means” (id. at 7 (citation omitted)).
Petitioner also seeks to exclude Exhibits F, G and their accompanying
testimony under FRE 403 because the “probative value of the evidence is
substantially outweighed by the risk of unfair prejudice or confusion of the
issues.” Id. at 8 (citation omitted). Petitioner explains that “[t]he real
question is not whether preliminary, single-term searches on generic search
engines would lead a POSITA to the reference; rather, the question is
‘whether an interested searcher would have been sufficiently capable of
finding the reference and examining its contents.’” Id. at 9 (citation
omitted).
Petitioner further seeks to exclude Exhibits F, G, and their
accompanying testimony under FRE 702 and 703. Id. at 10, 14. In support
of this position, Petitioner submits that “Mr. Smith did not establish
qualification as an expert in prior-art searching or in the field of optics” (id.
at 12) and “there is no evidence that experts in the field would reasonably
rely on truncated results from a single-query search to form an opinion about
the accessibility of Carlsohn” (id. at 14).
As to Exhibits F, G, and their accompanying testimony, Petitioner’s
motion is denied. As with Exhibit E and its accompanying testimony,
Petitioner’s arguments go to weight, not admissibility, and the Board is
IPR2020-00899
Patent 9,950,346 B2

106

capable of assessing the weight to be given the testimony without excluding
it.
Regarding relevance, and as pointed out by Patent Owner, Exhibits F,
G, and their accompanying testimony are relevant to rebut Dr. Arce’s
testimony pertaining to a POSITA’s use of Internet searches for accessing
Carlsohn. See Opp. 5 (citing Ex. 2022 ¶¶ 13–21); see also supra § II.D.3.
Specifically, Dr. Arce testified on cross examination that he would
“probably” use the Internet in searching for hyperspectral imaging. See
Ex. 2024, 68:15–20 (“Q. Would you use these same search terms if you
were a POSITA searching on the internet in hyperspectral imaging? A. If I
was a POSITA in this particular area? Q. Correct. In 2008. A. Probably,
yeah.”).
As to reliability, and as explained above, the policy considerations for
excluding expert testimony are less compelling in inter partes reviews than
in jury trials because the Board has significant experience in evaluating
expert testimony. See Nestle, IPR2015-00249, Paper 76 at 23.
Accordingly, we deny Petitioner’s motion to exclude Exhibits F, G,
and their accompanying testimony (paragraphs 15–21) of Mr. Smith’s
declaration (Ex. 2022).

3. Paragraphs 6, 25
Petitioner argues that Paragraphs 6 and 25 of Exhibit 2022 are
inadmissible as irrelevant under FRE 401 and unfairly prejudicial under FRE
403 “because a 2016 reference cannot establish the state of the art in 2007.”
See Mot. 7, 10; see also id. at 3 (summarizing Exhibit 2022 ¶¶ 6, 25 as
including “Testimony concerning the current state of the art with reference
to Geralyn Miller, Sorting it Out: How Imaging Systems Are Improving the
IPR2020-00899
Patent 9,950,346 B2

107

Way We Recycle, Possibility (Sept. 27, 2016)” (the “2016 Article”)); see
also Reply Opp. 7 (“Paragraphs 6 and 25 of Exhibit 2022 are inadmissible
under FRE 401 and 403 because Mr. Smith’s interpretation of a 2016
reference cannot establish the state of the art in 2007, and the risk of undue
prejudice outweighs their probative value.”).
Petitioner’s argument has merit. Mr. Smith appears to rely on the
2016 Article in his testimony at paragraphs 6 and 25 of Exhibit 2022. For
example, in paragraph 6 of Mr. Smith’s testimony, Mr. Smith quotes the
following statement from the 2016 Article:
Standard SWIR InGaAs cameras for plastic sorting have a
sensitivity up to 1.7μm, but hydrocarbon absorption bands are
more distinct between 1.9-2.4μm than below 1.7μm. This
requires hyperspectral mercury cadmium telluride (InGaAs
sensors can’t detect the subtle differences between office paper,
newspaper and cardboard. MCT) sensors, which are sensitive up
to 2.5μm. The InGaAs sensors used to sort plastics work well at
room temperature, but MCT sensors have to be cooled to -80°C,
making them more expensive. For sorting paper, their extended
hyperspectral range justifies the premium.
Ex. 2022 ¶ 6 & n.2 (providing a hyperlink to and quoting the 2016 Article).
As correctly argued by Petitioner, Mr. Smith’s citation to the 2016 Article is
not helpful in establishing the state of the art in 2007. See, e.g., Reply Opp.
7.
Nevertheless, we find that paragraphs 6 and 25 of Exhibit 2022 do not
rely exclusively on the 2016 Article, and include testimonial evidence that is
relevant to issues addressed in this Decision.
As to paragraph 6, Mr. Smith also cites to the Specification of the
’346 patent in testifying that a “POSITA reading the . . . claim limitation []
would understand that this would require at least some appropriate range in
IPR2020-00899
Patent 9,950,346 B2

108

order for identification to be accomplished.” Ex. 2022 ¶ 6 (citing Ex. 1001,
4:28–32). We find that this particular testimony is relevant in determining
an issue addressed in our Decision, specifically, whether the proper
construction of the claims requires a camera “configured to receive spectral
data up to at least 1700 nm,” as Patent Owner contends. See PO Resp. 6; see
also supra § II.C.1. Accordingly, we decline to exclude paragraph 6 of
Exhibit 2022 as irrelevant under FRE 401.
Regarding paragraph 25, this paragraph also includes testimony that
does not rely on the 2016 Article. For example, Mr. Smith testifies, “As
discussed at length in my first declaration, it is the differences between the
operation of Sinram and Carlsohn that would allow Sinram to be used at the
end of the conveyor belt, namely the difference in the amount of data that
must be possessed for a Sinram/LLA multiplexed technology verses
hyperspectral technology.” Ex. 2022 ¶ 25. This testimony is relevant to an
issue addressed in this Decision, namely, whether “a POSITA would not be
motivated by Sinram to move the Carlsohn hyperspectral identification
system over the second end of the belt,” as Patent Owner argues. See PO
Resp. 51; see also supra § II.H.3. Accordingly, we decline to exclude
paragraph 25 of Exhibit 2022 as irrelevant under FRE 401.
We also do not find that the probative value of the testimony
(Ex. 2022 ¶¶ 6, 25) is substantially outweighed by the danger of unfair
prejudice and confusion under FRE 403. As similarly explained above, the
danger of confusion and unfair prejudice in this proceeding is considerably
lower than in a conventional district court trial in front of a lay jury.
Accordingly, we deny Petitioner’s motion to exclude paragraphs 6 and
25 of Mr. Smith’s declaration (Ex. 2022).
IPR2020-00899
Patent 9,950,346 B2

109

B. Summary
We deny Petitioner’s Motion to Exclude.

IV. PATENT OWNER’S MOTION TO EXCLUDE
Patent Owner filed a Motion to Exclude Exhibit 1039. Paper 43.
We do not rely on Exhibit 1039 in this Final Written Decision.
Accordingly, we dismiss Patent Owner’s motion as moot. See Patent Trial
and Appeal Board Consolidated Trial Practice Guide 79–80 (Nov. 2019)
(“In the Board’s experience, consideration of the objected-to evidence is
often unnecessary to resolve the patentability of the challenged claims, and
the motion to exclude is moot.”).

IPR2020-00899
Patent 9,950,346 B2

110

V. CONCLUSION
Weighing the evidence of the disclosure of the references, the
competing testimony, and the reasoning to combine the references, we
determine that Petitioner has shown by a preponderance of the evidence that
claims 1–16 of the ’346 patent are unpatentable.

Claims 35 U.S.C. § Reference(s)/Basis
Claims
Shown
Unpatentable
Claims
Not Shown
Unpatentable
1–5, 7,
9–12
103 Carlsohn, Arleth 1–5, 7, 9–12
8 103 Carlsohn, Arleth,
Ulrichsen, Mayhak
8
1–7, 9–
16
103 Carlsohn, Arleth,
Sinram
1–7, 9–16
8 103 Carlsohn, Arleth,
Ulrichsen, Sinram
Mayhak
8
2 103 Carlsohn, Arleth,
Kampf
2
2  103  Carlsohn, Arleth,
Sinram, Kampf 
2   
12, 13 103 Carlsohn, Arleth,
Cowling2
12, 13
12, 13 103 Carlsohn, Arleth,
Sinram, Colwing2
12, 13
Overall
Outcome
1–16
IPR2020-00899
Patent 9,950,346 B2

111

VI. ORDER
Accordingly, it is:
ORDERED that claims 1–16 of the ’346 patent have been shown to be
unpatentable;
FURTHER ORDERED that Petitioner’s Motion to Exclude is denied;
FURTHER ORDERED that Patent Owner’s Motion to Exclude is
dismissed; and
FURTHER ORDERED that any party seeking judicial review must
comply with the notice and service requirements of 37 C.F.R. § 90.2.12

12 Should Patent Owner wish to pursue amendment of the challenged claims
in a reissue or reexamination proceeding subsequent to the issuance of this
Decision, we draw Patent Owner’s attention to the April 2019 Notice
Regarding Options for Amendments by Patent Owner Through Reissue or
Reexamination During a Pending AIA Trial Proceeding. See 84 Fed. Reg.
16,654 (Apr. 22, 2019). If Patent Owner chooses to file a reissue application
or a request for reexamination of the challenged patent, we remind Patent
Owner of its continuing obligation to notify the Board of any such related
matters in updated mandatory notices. See 37 C.F.R. § 42.8(a)(3), (b)(2).
IPR2020-00899
Patent 9,950,346 B2

112

For PETITIONER:

Peter Sullivan
Jeffrey Lewis
Stephen Kenny
FOLEY HOAG LLP
psullivan@foleyhoag.com
jidlewis@foleyhoag.com
skenny@foleyhoag.com

For PATENT OWNER:

Robert Brunelli
Patricia Ho
Kendria Pearson
SHERIDAN ROSS P.C.
rbrunelli@sheridanross.com
pho@sheridanross.com
kpearson@sheridanross.com