Detailed Action
Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
2. The Amendment filed 4/17/26 has been entered. Claims 38 and 41 have been cancelled. Claims 1, 4, 7, 9, 11, 13-14, 16-18, 21, 28, 39-40, 42-45 are pending.
3. The Information Disclosure Statements (IDS) filed 12/29/22 and 7/15/25 have been entered and acknowledged.
Claim Rejections - 35 USC § 112
4. The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
5. Claim 16 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention. Clam 16 lines 1-3 recite “a first output of the modified trained neural network-based decoder provides first reconstructed data corresponding to a first reconstruction from the latent without the modification” but this appears to contradict itself. It refers to the output of the modified trained neural network-based decoder but yet without the modification. It is not clear whether the output is coming from the decoder after the decoder has been modified or yet before the decoder has been modified, and thus the claim is vague and indefinite.
Claim Rejections - 35 USC § 103
5. The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
6. Claim(s) 1, 4, 7, 9, 14, 16, 17, 18, 21, 39-40, and 42-35 is/are rejected under 35 U.S.C. 103 as being unpatentable over Owada et al “Owada” (US 2020/0118003 A1) and Jung et al “Jung” (FR 3096538 A1) and Yada et al “Yada” (US 11010421 B2).
(Please see the attached copy of Yada and previously attached copy of Jung that number paragraphs in the same manner as that used in this Action).
7. Regarding claim 1, Owada shows: obtaining a latent representative of at least one part of at least one image (para 16, 24 show obtaining a latent image); using at least one update parameter representative of a modification to apply to a deep neural network-based decoder (para 25, 37, and 45 show the updated parameter used to modify the decoder network. Para 16, 22, and 42 show the decoder network is a deep neural network); modifying the deep neural network-based decoder based on the update parameter and providing a modified neural network based decoder (para 45, 48, 73 show updating/modifying the decoder based on the update parameter and providing the modified decoder for use; as noted, para 16, 22, and 42 show the decoder is a deep neural network); and reconstructing the at least one part of the at least one image from the latent using at least said modified deep neural network-based decoder (para 17, 24, 45, 73 show reconstructing the volume data to reconstruct the original image from the latent image, using the updated deep neural network-based decoder).
Owada not explicitly show decoding the update parameter itself and providing a decoded update parameter such that modifying the neural network is based on the decoded update parameter. Jung however does show decoding the update parameter providing a decoded update parameter, and that modifying the neural network is based on the decoded update parameter (para 35 shows decoding a control parameter from the coded data stream and providing it as useable data, and para 40 shows that control parameter corresponds to an update parameter of the neural network. Para 158, 159, 168, 202 show the neural network is used to decode image data). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention for the update parameter itself to be decoded in Owada, because it would provide an efficient way to use an encoder/decoder network to receive the data. Doing so allows the image and control data to be received together, decoded and used accordingly, as shown in Jung (para 32, 35, 36). Owada and Jung do not explicitly show the neural network decoder is already trained per se. Yada however does show using a trained neural network decoder to decode image data (para 34, 107, 121, 139 show providing the trained neural network decoder). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to use the trained neural network decoder in Owada, especially as modified by Jung, because it would provide an efficient way to fully utilize the decoder.
8. Regarding claim 4, Owada shows: a method, comprising obtaining at least one update parameter for modifying a deep-neural-network-based decoder (para 25, 37, 45 show receiving the updated parameters for modifying the decoder network. Para 16, 22, and 42 show the decoder network is a deep neural network) defined from a training of a deep neural network-based auto-encoder using a first training configuration (para 25, 27-28, 51 show the decoder is defined from the training of the encoder/decoder network [which per para 32 comprises the auto-encoder] using a particular training configuration), said at least one update parameter being obtained as a function of a training of said deep neural network-based auto-encoder using a second training configuration (para 25, 37, 39, 45 show the update parameter is based on the training of the encoder/decoder network using a second training configuration. Para 32, 41, 42, 66 show the function of the training of the encoder/decoder network using a second training configuration); encoding at least one part of at least one image using at least the neural network-based auto-encoder trained using the second training configuration (para 44, 66-67 then show after training the encoder/decoder network using the second training configuration, using the updated encoder of the encoder/decoder network to encode image data to output a latent image). Owada does not explicitly show encoding said at least one update parameter. Jung does show encoding the update parameter (two lines right before 140, para 140, 200 show encoding the updated parameters). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to encode the update parameter in Owada, because it would provide an efficient way to use an encoder/decoder network to transmit the data. Doing so allows the image and control data to be transmitted and received together, to eventually be decoded and used accordingly, as shown in Jung (para 32, 35, 36). Owada and Jung do not explicitly show the neural network decoder is already trained per se. Yada however does show using a trained neural network decoder to decode image data (para 34, 107, 121, 139 show providing the trained neural network decoder). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to use the trained neural network decoder in Owada, especially as modified by Jung, because it would provide an efficient way to fully utilize the decoder.
9. Claims 7 and 9 show the same features as claims 1 and 4 respectively, and are rejected for the same reasons. In addition, note Owada para 26-27 show the apparatus with processors configured to perform the method steps accordingly.
10. Regarding claim 14, in addition to that mentioned for claim 1, Owada shows said deep neural network-based decoder (which would be trained as explained for claim 1) comprises a hyper decoder configured for decoding side information used by a decoder configured for decoding said bitstream, and wherein modifying said deep neural network-based decoder comprises updating said hyper decoder (Owada para 25, 37, 45 show the hyper decoder with updated hyper parameters). Para 36 and 42 show the mean square error statistical side information that entropy decoders tend to focus on, but Owada does not explicitly mention that the information is used by an entropy decoder per se for entropy decoding the bitstream. Jung however does show the entropy decoder using side information to entropy decode the bitstream (para 29 show the entropy decoder using refinement datum and para 77 show the refinement data may include extracting statistical characteristics). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to use side information for an entropy decoder as is done in Jung, with the hyper decoder of Owada, because it would provide an efficient way to use a hyper decoder to decode the data for reconstructing an image. Doing so allows further refinement data to be decoded and used in reconstructing the image.
11. Regarding claim 16, in addition to that mentioned for claim 1, Owada shows a first output of the deep-neural-network-based decoder (which would be trained as explained for claim 1) which provides first reconstructed data obtained corresponding to reconstruction from the latent without the modification (para 16, 24, 40 show outputting the reconstructed data obtained by the deep neural network decoder before the updated modification), the first reconstructed data being used for reference by the modified (trained) deep-neural-network-based decoder (para 48-50 show the generated/reconstructed latent image may be used as an explosion image to serve as the reference image), and wherein a second output of the modified (trained) deep-neural-network-based decoder provides second reconstructed data corresponding to a second reconstruction from the latent with the modification, the second reconstructed data being used for display (para 50 shows the decoder outputs a second reconstructed latent image obtained with the modifications made to the decoder accordingly for display). Note also that although Owada does not explicitly mention the first reconstructed data used for reference in a predictive coding loop per se, nevertheless Jung shows the first reconstructed data used for reference in a predictive coding loop by the modified neural network decoder (para 31 shows the predictive data refinement based on the first/reference data). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to use this in Owada, because it would provide an efficient way to apply machine learning without having to require fully labeled datasets or every piece of raw data.
12. Regarding claim 17, in addition to that mentioned for claim 4, Owada shows obtaining said at least one update parameter comprises: training said deep neural network-based auto-encoder using said first training configuration (para 25, 27-28 show training the auto-encoder using a first training configuration); storing learnable parameters of a decoder of said deep neural network-based auto-encoder (para 25, 37, 45 show the updated parameters as a result of the neural network learning are stored and available), note also that as explained for claim 4, Yada para 34, 107, 121, 139 show providing the trained neural network based decoder; and said deep-neural-network-based decoder using said second training configuration, wherein said retraining comprises modifying said deep neural network-based decoder (which would be trained as explained for claim 4), said at least one update parameter being representative of said modification (para 45, 47, 48 show the decoder uses the second configuration based on the updated parameters which modify the decoder).
13. Regarding claim 18, in addition to that mentioned for claim 17, Owada shows the retraining comprises jointly retraining an encoder part of said deep-neural-network-based auto-encoder using said second training configuration (para 43, 44, 58, 66 show the encoder is also trained using the second configuration based on the updated parameters. Para 43 and Figure 5 show how the auto-encoder includes the encoder and decoder networks which are jointly being trained).
14. Regarding claim 21, in addition to that mentioned for claim 4, the second training configuration comprises a loss function based on at least one of a subjective quality and a metric for a machine task or wherein the second training configuration comprises a data set with specific video content type (note the alternative recitation. Owada para 36, 46, 47, 70 show updated training using the loss function based on a particular color content with a reference image. Para 17 shows the data may specifically be video data).
15. Claims 39-40 show the same features as claims 1 and 4 respectively, and are rejected for the same reasons. Additionally, note that para 27-28 shows the memories storing instructions to cause processors to carry out the method steps.
16. Claim(s) 11, 13, 28, and 42-45 is/are rejected under 35 U.S.C. 103 as being unpatentable over Owada and Jung and Yada and Karras et al “Karras” (US 2019/0171936 A1).
17. Regarding claim 11, in addition to that mentioned for claim 1, Owada and Jung and Yada do not explicitly show that modifying said deep neural network-based decoder comprises at least one of adding at least one new layer to said deep neural network-based decoder and updating at least one layer of a set of layers of said deep neural network. Please note the alternative recitation. Karras however does show modifying the deep neural network based decoder (which would be trained as explained for claim 1) comprises adding a new layer to the deep neural network-based decoder (para 49, 52, 58 shows the updated parameter adds a new layer to the neural network based encoder/decoder system. Para 70, 88 shows the new layer may be added to the decoder). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to have the update parameter add a new layer to the decoder in Owada, especially as modified by Jung and Yada, because it would provide an efficient way to train a decoder network to reconstruct the image. Doing so allows the neural network of the decoder to add a new layer to process the data with the new training configuration.
18. Regarding claim 13, in addition to that mentioned for claim 4, Owada and Jung and Yada do not explicitly show that modifying said deep neural network-based decoder comprises at least one of adding at least one new layer to said deep neural network-based decoder and updating at least one layer of a set of layers of said deep neural network. Please note the alternative recitation. Karras however does show modifying the deep neural network based decoder (which would be trained as explained for claim 1) comprises adding a new layer to the deep neural network-based decoder (para 49, 52, 58 shows the updated parameter adds a new layer to the neural network based encoder/decoder system. Para 70, 88 shows the new layer may be added to the decoder). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to have the update parameter add a new layer to the decoder in Owada, especially as modified by Jung and Yada, because it would provide an efficient way to train a decoder network to reconstruct the image. Doing so allows the neural network of the decoder to add a new layer to process the data with the new training configuration.
19. Regarding claim 28, in addition to that mentioned for claim 1, please note the alternative recitation. Owada and Jung and Yada do not explicitly show that the update parameter comprises an indication of adding at least one new layer to said deep neural network-based decoder. (Note though that the deep neural network-based decoder would be trained as explained for claim 1). Karras however does show an update parameter comprises adding a new layer to the deep neural network-based decoder (para 49, 52, 58 shows the updated parameter adds a new layer to the neural network based encoder/decoder system. Para 70, 88 shows the new layer may be added to the decoder). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to have the update parameter add a new layer to the decoder in Owada, especially as modified by Jung and Yada, because it would provide an efficient way to train a decoder network to reconstruct the image. Doing so allows the neural network of the decoder to add a new layer to process the data with the new training configuration.
20. Regarding claim 42, in addition to that mentioned for claim 4, please note the alternative recitation. Owada and Jung and Yada do not explicitly show that the update parameter comprises an indication of adding at least one new layer to said deep neural network-based decoder. (Note though that the deep neural network-based decoder would be trained as explained for claim 1). Karras however does show an update parameter comprises adding a new layer to the deep neural network-based decoder (para 49, 52, 58 shows the updated parameter adds a new layer to the neural network based encoder/decoder system. Para 70, 88 shows the new layer may be added to the decoder). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to have the update parameter add a new layer to the decoder in Owada, especially as modified by Jung and Yada, because it would provide an efficient way to train a decoder network to reconstruct the image. Doing so allows the neural network of the decoder to add a new layer to process the data with the new training configuration.
21. Regarding claim 43, in addition to that mentioned for claim 7, please note the alternative recitation. Owada and Jung and Yada do not explicitly show that the update parameter comprises an indication of adding at least one new layer to said deep neural network-based decoder. (Note though that the deep neural network-based decoder would be trained as explained for claim 1). Karras however does show an update parameter comprises adding a new layer to the deep neural network-based decoder (para 49, 52, 58 shows the updated parameter adds a new layer to the neural network based encoder/decoder system. Para 70, 88 shows the new layer may be added to the decoder). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to have the update parameter add a new layer to the decoder in Owada, especially as modified by Jung and Yada, because it would provide an efficient way to train a decoder network to reconstruct the image. Doing so allows the neural network of the decoder to add a new layer to process the data with the new training configuration.
22. Regarding claim 44, in addition to that mentioned for claim 9, please note the alternative recitation. Owada and Jung and Yada do not explicitly show that the update parameter comprises an indication of adding at least one new layer to said deep neural network-based decoder. (Note though that the deep neural network-based decoder would be trained as explained for claim 1). Karras however does show an update parameter comprises adding a new layer to the deep neural network-based decoder (para 49, 52, 58 shows the updated parameter adds a new layer to the neural network based encoder/decoder system. Para 70, 88 shows the new layer may be added to the decoder). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to have the update parameter add a new layer to the decoder in Owada, especially as modified by Jung and Yada, because it would provide an efficient way to train a decoder network to reconstruct the image. Doing so allows the neural network of the decoder to add a new layer to process the data with the new training configuration.
23. Regarding claim 45, in addition to that mentioned for claim 39, please note the alternative recitation. Owada and Jung and Yada do not explicitly show that the update parameter comprises an indication of adding at least one new layer to said deep neural network-based decoder. (Note though that the deep neural network-based decoder would be trained as explained for claim 1). Karras however does show an update parameter comprises adding a new layer to the deep neural network-based decoder (para 49, 52, 58 shows the updated parameter adds a new layer to the neural network based encoder/decoder system. Para 70, 88 shows the new layer may be added to the decoder). It would have been obvious to a person with ordinary skill in the art before the effective date of the claimed invention to have the update parameter add a new layer to the decoder in Owada, especially as modified by Jung and Yada, because it would provide an efficient way to train a decoder network to reconstruct the image. Doing so allows the neural network of the decoder to add a new layer to process the data with the new training configuration.
24. Applicant's arguments filed 4/17/26 have been fully considered but they are not persuasive. Applicant argues that Owada and Jung do not show the trained neural network based decoder, but Yada is brought in to show this. Furthermore, regarding claim 9, Applicant may recite a first and second training configurations, but really the difference is simply that the first “configuration” is before obtaining the updated parameter and the second “configuration” is after obtaining the updated parameter. Other aspects of the training process may be the same. As such, this is shown in Owada as explained in the Action. Likewise regarding claim 16, although it recites a first and second reconstruction, really the first reconstruction simply pertains to before receiving the modification from the updated parameter, and the second reconstruction pertains to after receiving the modification from the updated parameter, as explained in the claim language itself. Other aspects of the reconstruction process may be the same and as such, this is shown in Owada as explained in the Action. The newly added language of the predictive coding loop however is shown in Jung, as explained in the Action.
Conclusion
25. Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
26. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
a) Sun (CN 110569961 A) encodes/decodes parameters used for image reconstruction.
b) Denli (CA 3122686) trains deep neural networks for image decoding.
c) Kim (US 20210295606) reconstructs latent image data.
27. Any inquiry concerning this communication or earlier communications from the examiner should be directed to STEVEN PAUL SAX whose telephone number is (571)272-4072. The examiner can normally be reached Monday - Friday, 9:30 - 6:00 Est.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Usmaan Saeed, can be reached at 571-272-4046. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/STEVEN P SAX/Primary Examiner, Art Unit 2146
Prosecution Insights