General Electric Company

Patent Trials and Appeals Board

Jul 28, 2021

2020003824 (P.T.A.B. Jul. 28, 2021)

UNITED STATES PATENT AND TRADEMARK OFFICE
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
15/400,763 01/06/2017 Lishui Cheng 316189-US-1
(GEMS:0660)
2047
68174 7590 07/28/2021
GE HEALTHCARE
c/o FLETCHER YODER, PC
P.O. BOX 692289
HOUSTON, TX 77269-2289
EXAMINER
ANSARI, TAHMINA N
ART UNIT PAPER NUMBER
2662
NOTIFICATION DATE DELIVERY MODE
07/28/2021 ELECTRONIC
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address(es):
docket@fyiplaw.com
rariden@fyiplaw.com
robinson@fyiplaw.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte LISHUI CHENG, BRUNO KRISTIAAN BERNARD DE MAN,
SHESHADRI THIRUVENKADAM, SANGTAE AHN,
LIN FU, and HAO LAI
Appeal 2020-003824
Application 15/400,763
Technology Center 2600

Before JAMES B. ARPIN, HUNG H. BUI, and
AMBER L. HAGY, Administrative Patent Judges.

HAGY, Administrative Patent Judge.

DECISION ON APPEAL
STATEMENT OF THE CASE
Pursuant to 35 U.S.C. § 134(a), Appellant1 appeals from the
Examiner’s decision to reject claims 1–21, all of the pending claims.
See Final Act. 1–2; Appeal Br. 2. We have jurisdiction under
35 U.S.C. § 6(b).
We affirm.

1 Appellant refers to “applicant” as defined in 37 C.F.R. § 1.42. Appellant
identifies the real party in interest as General Electric Company. Appeal
Br. 1.
Appeal 2020-003824
Application 15/400,763

2
CLAIMED SUBJECT MATTER
The subject matter of the present application “relates to tomographic
reconstruction, and in particular to the use of deep learning techniques to
accelerate iterative reconstruction approaches.” Spec. ¶ 1. By way of
background, the Specification describes “[n]on-invasive imaging
technologies” that “allow images of the internal structures or features of a
patient/object to be obtained without performing an invasive procedure on
the patient/object.” Id. ¶ 2. The Specification further describes that “there is
a need for reconstruction techniques that may provide improved benefits,
such as increased reconstruction efficiency or speed, while still achieving
good image quality or allowing a low patient dose.” Id. ¶ 3. The
Specification purports to describe a learning model that is trained to
reconstruct image signals using an artificial neural network. See, e.g., id.
¶¶ 4–6.
Claims 1, 6, and 17 are independent. Claim 1, reproduced below, is
representative:
1. A neural network training method, comprising:
acquiring, via data acquisition circuitry of a medical
imaging system, a plurality of sets of X-ray scan data;
performing, via a processor of the medical imaging
system, an iterative reconstruction of each set of X-ray scan data
to generate one or more input images and one or more target
images for each set of X-ray scan data, wherein the one or more
input images correspond to lower iteration steps or earlier
convergence status of the iterative reconstruction than the one or
more target images;
training, via the processor, a neural network to generate a
trained neural network by providing the one or more input
Appeal 2020-003824
Application 15/400,763

3
images and corresponding one or more target images for each set
of X-ray scan data to the neural network; and
utilizing, via the processor, the trained neural network to
generate a reconstructed X-ray image.
REFERENCES
The Examiner relies on the following references:
Name Reference Date
Balan2 US 2005/0196065 A1 Sept. 8, 2005
Chen US 2015/0086097 A1 Mar. 26, 2015
REJECTION
Claims 1–21 stand rejected under 35 U.S.C. § 1033 as obvious over
the combined teachings of Chen and Balan. Final Act. 5–32.
OPINION
We have considered Appellant’s arguments (Appeal Br. 5–11; Reply
Br. 2–6) in light of the Examiner’s findings and explanations (Final Act. 2–
32; Ans. 3–21). For the reasons set forth below, we are not persuaded of
Examiner error in the rejection of the pending claims, and we, therefore,
sustain the Examiner’s rejection.
Appellant argues only claim 1 with particularity, and incorporates by
reference those arguments with regard to independent claims 6 and 17 and
the respective dependent claims. See Appeal Br. 6–10. Therefore, based on

2 All references are cited using the first-named inventor.
3 The Examiner’s rejection is under the provisions of Title 35 of the United
States Code in effect after the effective date of the Leahy-Smith America
Invents Act of 2011.
Appeal 2020-003824
Application 15/400,763

4
Appellant’s arguments, we decide the appeal of claims 1–21 based on claim
1 alone. See 37 C.F.R. § 41.37(c)(1)(iv) (2019).
The Examiner finds Chen teaches most of the limitations of claim 1,
including acquiring X-ray scan data, performing iterative reconstruction, and
generating a reconstructed X-ray image. Final Act. 5–10 (citing Chen
¶¶ 15–27, 29, 50–54, 64). The Examiner finds, however, that Chen “does
not specifically teach: ‘a neural network,’” for which the Examiner relies on
Balan’s teachings in combination with Chen’s. See id. at 10–15 (citing
Balan Abs., ¶¶ 11, 19–20, 26, 44, 67–73, 79–92).
Appellant argues that the Examiner’s findings are in error because
Chen does not teach a neural network, and, therefore, “Chen fails to teach or
suggest utilizing input images (corresponding to lower iteration steps or an
earlier convergence status) and corresponding target images as inputs to a
neural network for training,” and also “fails to teach or suggest utilizing a
trained neural network to generate a reconstructed X-ray image.” Appeal
Br. 7–8; Reply Br. 2, 4. Appellant acknowledges Balan teaches a neural
network, but argues that Balan “does not teach or suggest utilizing the neural
network for image processing, much less medical imaging processing of X-
ray scan data.” Appeal Br. 8; Reply Br. 2–4. According to Appellant, Balan
discloses only “a neural network trained on speech and noise signals and
utilized for enhancing a signal.” Appeal Br. 8 (emphasis omitted); Reply
Br. 2–3.
We are not persuaded of Examiner error. Appellant raises several
challenges, in piecemeal and conclusory fashion, to the Examiner’s reliance
on each of the cited references. As an initial matter, we note that such
piecemeal challenge is inherently flawed. The test for obviousness is not
Appeal 2020-003824
Application 15/400,763

5
whether the claimed invention must be expressly taught or suggested in any
one or all of the references. “Rather, the test is what the combined teachings
of the references would have suggested to those of ordinary skill in the art.”
In re Keller, 642 F.2d 413, 425 (CCPA 1981) (citations omitted) (emphasis
added). Thus, where, as here, the rejections are based upon the combined
teachings of the references, “[n]on-obviousness cannot be established by
attacking references individually.” In re Merck & Co., Inc., 800 F.2d 1091,
1097 (Fed. Cir. 1986) (citing Keller, 642 F.2d at 425). In addition, a
reference “must be read, not in isolation, but for what it fairly teaches in
combination with the prior art as a whole.” Id.
In addition, we find that Appellant’s individual challenges are
conclusory and/or not well-founded factually, and, hence, are not persuasive
of Examiner error. With regard to the Examiner’s reliance on Chen,
Appellant argues Chen is “silent with regard to input images
(corresponding to lower iteration steps or an earlier convergence status) and
corresponding target images as inputs for training (via a neural network or
any other structure).” Appeal Br. 9; Reply Br. 2. We disagree with
Appellant regarding Chen’s silence pertaining to input and target images.
As the Examiner finds, Chen “teaches the use of a source image, and
that a target image of an underlying object is reconstructed by using ordered
subsets of data that converges from an initial estimate is updated,” wherein
the “initial input is a source image.” Ans. 17 (citing Chen ¶¶ 40–42). The
Examiner also finds that “Chen further teaches that this target image[] serves
as inputs for iterative reconstruction.” Id. (citing Chen ¶ 47). The
Examiner’s findings are supported by the cited disclosures, and we adopt
Appeal 2020-003824
Application 15/400,763

6
them as our own. Appellant’s conclusory contentions do not address the
Examiner’s findings and do not show Examiner error.
With regard to the Examiner’s reliance on Balan, Appellant argues
Balan fails to disclose a reconstruction process applicable to image signals.
See Appeal Br. 8. We disagree with Appellant.
As the Examiner finds and Appellant concedes, Balan “teaches the
use of a neural network for signal data, but mischaracterizes the teachings of
Balan when indicating that it is exclusively for speech and noise signals.”
Ans. 19–20. The Examiner further finds that Balan “teaches the use of
image signals” (id. at 20 (citing Balan ¶¶ 27, 39)), and “teaches nonlinear
signal reconstruction that does not use a noise component of the signal” (id.
(citing Balan ¶¶ 88–92, Fig. 3)). Thus, “Balan’s teaching[] for signal
reconstruction lend itself well for combination with the teachings of Chen,
and can easily be applied to any image data including X-ray image data.”
Id. The Examiner’s findings are supported by the cited teachings, and we
adopt them as our own. Appellant’s conclusory arguments do not address
these findings, are contradicted by Balan’s disclosures, and, thus, do not
show Examiner error.
With regard to the reason for the combination, the Examiner finds:
It would have been obvious before the effective filing date of the
claimed invention to one of ordinary skill in the art to modify
Chen’s methods and systems for iterative reconstruction of
medical image data to use Balan’s algorithms for signal
enhancement as they are both in the field of image analysis and
processing. The determination of obviousness is predicated
upon the following findings: One skilled in the art would have
been motivated to modify Chen’s iterative image reconstruction
in this manner in order to enhance the image signal and reduce
any noise with a learning system such as a neural network.
Appeal 2020-003824
Application 15/400,763

7
Furthermore, the prior art collectively includes each element
claimed (though not all in the same reference), and one of
ordinary in the art could have combined the elements in the
manner explained above using known engineering design,
interface and programming techniques, as they are both directed
towards the processing of image signal data. Chen
fundamentally teaches the claimed features, but does not rely on
a neural network; the modification to add the neural network
proposed by Balan could be made, without changing the
“fundamental” operating principle of Chen, while the teaching of
Balan continues to perform the same function as originally taught
prior to being combined, in order to produce the repeatable and
predictable result of estimate a noise-free part of the linear
transformed signal and to use a learned system to enhance the
overall reconstructive process.
Id. at 15–16; see also id. at 20. The Examiner’s findings are supported by
the disclosures of the cited references, and we adopt them as our own.
For the foregoing reasons, we are not persuaded of Examiner error in
the rejection of claim 1, and we, therefore, sustain that rejection, along with
the rejection of independent claims 6 and 17, and all of the dependent
claims, which are all argued collectively with claim 1. See Appeal Br. 5–11.
CONCLUSION
The Examiner’s obviousness rejection of claims 1–21 is sustained.

Appeal 2020-003824
Application 15/400,763

8
DECISION SUMMARY

In summary:
Claim(s)
Rejected
35 U.S.C. § Reference(s)/Basis Affirmed Reversed
1–21 103 Chen, Balan 1–21

TIME PERIOD FOR RESPONSE
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a). See 37 C.F.R.
§ 1.136(a)(1)(iv).
AFFIRMED