DETAILED ACTION
This Office action for U.S. Patent Application No. 18/410,581 is responsive to communications filed 28 January 2026, in reply to the Non-Final Rejection of 31 October 2025.
Claims 1–29 are pending.
In the previous Office action, claims 3, 9, 15, and 24 were rejected under 35 U.S.C. § 112(b) as indefinite. Claims 1–4, 8–16, 18, and 23–27 were rejected under 35 U.S.C. § 102(a)(1) as anticipated by Y. Shi, Y. Ge, J. Wang, & J. Mao, “Alpha VC: High-Performance and Efficient Learned Video Compression”, 17 Proc. of European Conf. on Computer Vision (Oct. 2022) (“Shi”). Claims 5 and 17 were rejected under 35 U.S.C. § 103 as obvious over Shi in view of A. Mercat, M. Viitanen, & J. Vanne, “UVG Dataset: 50/120fps 4k Sequence for Video Codec Analysis and Development” (Tampere U., 2020) (“Mercat”). Claims 6, 7, 19, and 20 were rejected under 35 U.S.C. § 103 as obvious over Shi in view of T. Ladune, P. Philippe, W. Hamidouche, L. Zhang, & O. Déforges, “Conditional Coding for Flexible Learned Video Compression”, 28 April 2021 (“Ladune”).
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendments
Applicant’s amendments to the claims have been considered. The rejections of claims 3, 9, 15, and 24 under 35 U.S.C. § 112(b) are withdrawn. The amendments and the accompanying argument (28 January 2026 “REMARKS” (“Rem.”) at p. 6) clarify that the claims’ intent is to make the threshold being open-ended and closed-ended as two options.
Response to Arguments
Applicant's arguments filed with respect to claim 1 have been fully considered but they are not persuasive.
With respect to the allegation that in Shi a conditional I-frame does not have accumulated errors (Rem. 7), while Shi does teach that a conditional I frame mitigates accumulated error relative to another P frame, Shi § 3.4 specifies this occurs in a rate-distortion tradeoff, with the cI frames having distortion. The conditional I frames are not new reference frames, but are still coded relative to a previous I-frame. § 3.1. Shi, published in 2022, does not simply assign intra frames adaptively, something known since at least 2006, but is a different kind of frame with its entropy coding dependent on a previous P frame.
With respect to the allegation that Shi does not teach generated a conditional I frame according to either claimed method of selecting a corrupt I-frame (Rem. 7–8), first, these methods are claimed alternatively and as two options. Under the Broadest Reasonable Interpretation standard, it is only required for one of the PSNR ratio and inserted errors to be present in the prior art for the claim to be anticipated. Specifically considering the PSNR threshold limitation, Also, the limitation as written of the corrupt I-frame having a PSNR that meets a threshold is broad enough to encompass the conditional I-frame being selected to avoid the PSNR of the next P-frame from being too high, as in Shi. Specifically concerning the error insertion option, claim 1 does not teach a specific method of inserting the errors. Compare with claim 2 as amended, which recites inserting the errors into an already generated I-frame. Claim 1, in contrast, is broader, and can encompass inserting the errors in the I-frame organically as it is being generated.
With respect to the allegation that Shi does not teach training a P-frame model, the claim does not require training a neural network to improve all future coding by the specific implementation of the coder, but only training a generic “P-frame model”. This can include the P-frames that follow the specific corrupt I-frame before the next I-frame. In Shi, presenting better P-frames that depend from the conditional I-frame, and producing a conditional I-frame that can be used as a reference frame for better P-frames, is within the scope of the very broad claim as originally filed. It is suggested that Applicant amend the claim to recite more specific limitations for training a neural media coder as positive steps, the preamble having questionable patentable weight under United States practice.
Applicant’s arguments with respect to claim 2 have been considered but are moot in view of new grounds of rejection. It is respectfully submitted that US 2022/0312017 A1 (“Bovik”) teaches the material of claims 2 and 14 as amended.
Applicant argues against the rejection of claims 5 and 17 under 35 U.S.C. § 103 separately, alleging the Mercat video clips are used for testing, not for training. However, Mercat was only cited for the existence of pre-configured video clips having known PSNR values, not for using them for training. The claimed training the P-frame model was given in claim 1, rejected over Shi alone, not in combination with Mercat. Applicant is reminded that one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 U.S.P.Q. 871 (C.C.P.A. 1981); In re Merck & Co., 800 F.2d 1091, 231 U.S.P.Q. 375 (Fed. Cir. 1986).
Claim Rejections - 35 U.S.C. § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. §§ 102 and 103 (or as subject to pre-AIA 35 U.S.C. §§ 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. § 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1, 3, 4, 8–13, 15, 16, 18, and 23–27 are rejected under 35 U.S.C. §102(a)(1) as being anticipated by Yibo Shi, Yunying Ge, Jing Wang, & Jue Mao, “AlphaVC: High-Performance and Efficient Learned Video Compression”, 17 Proc. of European Conf. on Computer Vision (Oct. 2022) (“Shi”)1.
Shi, directed to video compression, teaches with respect to claim 1 a method of training a P-frame model for a neural media coder (§ 2.2, learning-based video compression), the method comprising:
acquiring a corrupt I-frame (§§ 3.1, 3.4; conditional I-frame that has accumulated error relative to the first I-frame), the corrupt I-frame having at least one of
a) peak signal-to-noise ratio that meets a threshold (§ 3.1, conditional I-frame selected instead of P-frame based on accumulated error being over a threshold to prevent further error; §§ 3.4, 4,2, rate-distortion being measured with PSNR) or
b) one or more areas of inserted errors (§§ 3.1, 3.4; accumulated errors) and
training the P-frame model using the corrupt I-frame (§ 3.1, reference frames including conditional I-frames “participate in and affect the reconstruction of” subsequent P-frames).
Regarding claim 3, Shi teaches the method of claim 1, wherein the corrupt I-frame has the PSNR that meets the threshold[,] and wherein to meet the threshold, the PSNR is lower than the threshold or the PSNR is lower than or equal to the threshold (§ 3.1, conditional I-frame selected to stabilize entropy if accumulated errors would prevent another P frame from having an acceptable quality).
Regarding claim 4, Shi teaches the method of claim 1, wherein the threshold is based on a Lagrange multiplier associated with the P-frame model (§ 3.4, Lagrangian multiplier of distortion to determine the tradeoff between bit rate and error).
Regarding claim 8, Shi teaches a method of coding media data, the method comprising:
applying a pre-trained P-frame model to the media data (§§ 3.1–3.3, P-frame coding),
the pre-trained P-frame model being trained using a corrupt I-frame (§ 3.1, reference frames including conditional I-frames that have accumulated errors “participated in and affect the reconstruction of” subsequent P-frames),
the corrupt I-frame (§§ 3.1, 3.4; conditional I-frame that has accumulated error relative to the first I-frame) having at least one of
a) a peak signal-to-noise ratio (PSNR) that meets a threshold (§ 3.1, conditional I-frame selected based on accumulated error being over a threshold; §§ 3.4, 4.2, rate-distortion being measured with PSNR) or
b) one or more areas of inserted errors (§§ 3.1, 3.4; accumulated errors); and
coding the media data based on the application of the pre-trained P-frame model to the media data (Fig. 2, video compression scheme including P-frames).
Regarding claim 9, Shi teaches the method of claim 8, wherein the corrupt I-frame has the PSNR that meets the threshold, and wherein to meet the threshold, the PSNR is lower than the threshold or the PSNR is lower than or equal to the threshold (§ 3.1, conditional I-frame selected to stabilize entropy if accumulated errors would prevent another P frame from having an acceptable quality).
Regarding claim 10, Shi teaches the method of claim 8, where media data comprises video data (Fig. 2, compressing the I-frames and P-frames themselves).
Regarding claim 11, Shi teaches the method of claim 8, wherein coding comprises encoding (Fig. 2, compression).
Regarding claim 12, Shi teaches the method of claim 8, wherein coding comprises decoding (§§ 2, 3.1; codec includes compression and decoding).
Regarding claim 13, Shi teaches a device comprising:
one or more memories configured to store media data and a P-frame model (§§ 3.3, 4.2; operation on CPU or GPU); and
one or more processors implemented in circuitry and coupled to the one or more memories (§§ 3.3, 4.2; operation on CPU or GPU), the one or more processors being configured to:
acquire a corrupt I-frame (§§ 3.1, 3.4; conditional I-frame that has accumulated error relative to the first I-frame), the corrupt I-frame having at least one of
a) a peak signal-to-noise ratio (PSNR) lower than a threshold (§ 3.1, conditional I-frame selected based on accumulated error being over a threshold; §§ 3.4, 4.2, rate-distortion being measured with PSNR) or
b) one or more areas of inserted errors (§§ 3.1, 3.4; accumulated errors); and
train the P-frame model using the corrupt I-frame (§ 3.1, reference frames including conditional I-frames “participated in and affect the reconstruction of” subsequent P-frames).
Regarding claim 15, Shi teaches the device of claim 13, wherein the corrupt I-frame has the PSNR that meets the threshold and wherein to meet the threshold, the PSNR is lower than the threshold or lower than or equal to the threshold (§ 3.1, conditional I-frame selected to stabilize entropy if accumulated errors would prevent another P frame from having an acceptable quality).
Regarding claim 16, Shi teaches the device of claim 13, wherein the threshold is based on a Lagrange multiplier associated with the P-frame model (§ 3.4, Lagrangian multiplier of distortion to determine the tradeoff between bit rate and error).
Regarding claim 18, Shi teaches the device of claim 13, wherein the threshold comprises a Lagrange multiplier (§ 3.4, Lagrangian multiplier of distortion to determine the tradeoff between bit rate and error).
Regarding claim 21, Shi teaches the device of claim 13, further comprising a camera configured to capture the media data (Fig. 3, current frame input; Fig. 7a, “original image” input into the codec).
Regarding claim 22, Shi teaches the device of claim 13, further comprising a display configured to display the media data (passim, reconstructed frames).
Regarding claim 23, Shi teaches a device for coding media data, the device comprising:
one or more memories configured to store media data (§§ 3.3, 4.2; operation on CPU or GPU); and
one or more processors implemented in circuitry and coupled to the one or more memories (id.), the one or more processors being configured to:
[perform the claim 8 method] (claim 8 rejection supra).
Regarding claim 24, Shi teaches the device of claim 23, wherein the corrupt I-frame has the PSNR that meets the threshold and wherein to meet the threshold, the PSNR is lower than the threshold or equal to the threshold (§ 3.1, conditional I-frame selected to stabilize entropy if accumulated errors would prevent another P frame from having an acceptable quality).
Regarding claim 25, Shi teaches the device of claim 23, where media data comprises video data (Fig. 2, compressing the I-frames and P-frames themselves).
Regarding claim 26, Shi teaches the device of claim 23, wherein as part of coding the media data, the one or more processors are configured to encode the media data (Fig. 2, compression).
Regarding claim 27, Shi teaches the device of claim 23, wherein as part of coding the media data, the one or more processors are configured to decode the media data (§§ 2, 3.1; codec includes compression and decoding).
Regarding claim 28, Shi teaches the device of claim 23, further comprising a camera configured to capture the media data (Fig. 7a, “original image” input into the codec).
Regarding claim 29, Shi teaches the device of claim 23, further comprising a camera configured to capture the media data (Fig. 3, current frame input; Fig. 7a, “original image” input into the codec).
Claim Rejections - 35 U.S.C. § 103
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.
Claims 2 and 14 are rejected under 35 U.S.C. § 103 as obvious over Shi in view of European Patent Application Publication No. EP 3799431 A1 (Chadha”).
Claims 2 and 14, which appear to be the most representative of the intended invention, recite a method of generating the corrupt I-frame by first generating an I-frame and then inserting errors into the I-frame. This differs from Shi in which the conditional I-frames are produced to prevent the organic accumulation of errors in a GOP from propagating further by continuing to produce P-frames.
Regarding claim 2, Chadha, directed to video compression, teaches with respect to claim 2 the method of claim 1, further comprising generating the corrupt I-frame, wherein generating the corrupt I-frame comprises:
generating an I-frame (¶ 0024, transforming pixel input data according to typical video coding operations); and
inserting errors into the I-frame to generate the corrupt I-frame (¶ 0027, corrupting the transformed pixel data to emulate various coding artifacts).
It would have been obvious to one of ordinary skill in the art at the time of effective filing to generate the conditional I-frames according to a set corruption or degrading level, as taught by Chadha, in order to prepare the encoder for an upcoming loss of available bandwidth. Chadha ¶¶ 0028, 0059–61.
Regarding claim 14, Shi in view of Chadha teaches the device of claim 13, wherein the one or more processors are further configured to generate the corrupt I-frame,
wherein as part of generating the corrupt I-frame, the one or more processors are configured to:
generate an I-frame (Chadha ¶ 0024, transforming pixel input data according to typical video coding operations); and
insert errors into the I-frame to generate the corrupt I-frame (¶ 0027, corrupting the transformed pixel data to emulate various coding artifacts).
Claims 5 and 17 are rejected under 35 U.S.C. § 103 as obvious over Shi in view of A. Mercat, M. Viitanen, & J. Vanne, “UVG Dataset: 50/120fps 4k Sequence for Video Codec Analysis and Development” (Tampere U., 2020) (“Mercat”)2.
Claims 5 and 17 are directed to pre-configuring the corrupt I-frame to meet a PSNR threshold. Shi does not teach this limitation. However, Mercat, a survey of a set of 16 standard video clips, teaches with respect to claims 5 and 17: wherein the corrupt I-frame is pre-configured to meet the threshold (Table 3, known rate-distortion characteristics). It would have been obvious to one of ordinary skill in the art at the time of effective filing to use a standard set of video sequences to train the Shi model, as taught by Mercat, in order to provide a known benchmark for codec performance.
Claims 6, 7, 19, and 20 are rejected under 35 U.S.C. § 103 as obvious over Shi in view of T. Ladune, P. Philippe, W. Hamidouche, L. Zhang, & O. Déforges, “Conditional Coding for Flexible Learned Video Compression”, (28 April 2021) (“Ladune”)3.
Claims 6, 7, 19, and 20 are directed to training the P-frame model with specific numbers of frames, with claims 6 and 19 limiting the model to no more than seven frames and claims 7 and 20 requiring three frames. The ellipses in Shi figure 2 preclude judgment on a count of frames in a Group of Pictures. However, Ladune, directed to conditional coding for video compression, teaches with respect to claims 6 and 19 training a P-frame model for less than or equal to seven frames (Fig. 1, § 3; three inter-coded frames between each P-frame following the first I frame analogous to the Shi conditional I-frames).
It would have been obvious to one of ordinary skill in the art at the time of effective filing to limit the number of inter-coded frames between reference frames, as taught by Ladune, in order to minimize rate-distortion loss. Ladune § 3.
Regarding claims 7 and 20, Shi in view of Ladune teaches training the P-frame model for three frames (Ladune Fig. 1, § 3; three inter-coded frames between each P-frame following the first I frame analogous to the Shi conditional I-frames).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
CN 116781912 A
US 2022/0312017 A1
The following prior art was found using an Artificial Intelligence assisted search using an internal AI tool that uses the classification of the application under the Cooperative Patent Classification (CPC) system, as well as from the specification, including the claims and abstract, of the application as contextual information. The documents are ranked from most to least relevant. Where possible, English-language equivalents are given, and redundant results within the same patent families are eliminated. See “New Artificial Intelligence Functionality in PE2E Search”, 1504 OG 359 (15 November 2022), “Automated Search Pilot Program”, 90 F.R. 48,161 (8 October 2025).
WO 2022/213500 A1
US 2021/0092449 A1
US 2020/0221099 A1
CN 117083857 A
WO 2024/010672 A1
US 2021/0211685 A1
Applicant's amendment necessitated the new ground of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See M.P.E.P. § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 C.F.R. § 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 C.F.R. § 1.17(a)) pursuant to 37 C.F.R. § 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.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to David N Werner whose telephone number is (571)272-9662. The examiner can normally be reached M--F 7:30--4:00 Central.
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, Dave Czekaj can be reached at 571.272.7327. 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.
/David N Werner/Primary Examiner, Art Unit 2487
1 This reference was cited as an ‘X’ reference in the International Search Report for corresponding International Application No. PCT/UW2025/010777, and was listed in the 6 August 2025 Information Disclosure Statement.
2 This reference was listed in the 30 April 2024 Information Disclosure Statement.
3 This reference was listed in the 30 April 2024 Information Disclosure Statement.