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 . 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 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. There are a total of 20 claims and claims 1-20 are pending.
Response to Amendment
Applicant's argument, filed on December 22,2025 has been entered and carefully considered. Claims 1, 12-16, 19 and 20 are amended and claims 1-20 are pending.
Response to Arguments
Applicant's arguments filed on 12/22/2025 remarks have been fully considered but are moot in view of the new ground(s) of rejection which is deemed appropriate to address all of the needs at this time.
Claim Rejections - 35 USC § 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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over MA et al. (WO 2022072659 A1) in view of Chen et al.(NPL- AHGll: An Improved Unet-Based In-Loop Filter Method; pub. April 20-29, 2022; given by applicant in the IDS ) and further in view of Kang et al.( KR 20230125733 A).
Regarding claim 1, MA discloses a method of processing video data, the method comprising([abstract]-video data): in-loop filtering a current block of the video data using a neural network-based in-loop filter to generate an in-loop filtered current block([abstract and para [009-0010]]- a method of decoding video data, including: reconstructing, from a video bitstream, a picture frame that includes a luma component, a first and a second chroma components, and applying a trained neural network based in-loop filter to the reconstructed picture frame).
However, MA does not explicitly disclose wherein the neural network based in-loop filter is trained using an architecture comprising a U-Net architecture comprising one or more residual blocks and one or more transform blocks; and outputting the in-loop filtered current block.
In an analogous art, Chen discloses wherein the neural network based in-loop filter is trained using an architecture comprising a U-Net architecture comprising one or more residual blocks and one or more transformer blocks([see in Fig. 1, abstract and see in Section 2.2]- a convolutional neural network-based in-loop filtering method with QP based models. In this contribution, we modify the training strategy and Unet-Based model); and outputting the in-loop filtered current block([see in Fig. 1]-in Fig. 1 illustrates output). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to provide the technique of Chen to the modified system of MA a convolutional neural network-based in-loop filtering method with QP based models to modify the training strategy and Unet-Based model from JVET-Y0086 for the inter frame to improve the compression efficiency in JVET-Y0086 [Chen; abstract/Introduction].
However, the combination of MA and Chen don’t exclusive discloses wherein the neural network based in-loop filter is trained using an architecture comprising a U-Net architecture comprising one or more residual blocks and one or more transformer blocks.
In an analogous art, Kang discloses wherein the neural network based in-loop filter is trained using an architecture comprising a U-Net architecture comprising one or more residual blocks and one or more transformer blocks([para 0168]-the transformer-based in-loop filter may be a fixed-coefficient in-loop filter. Therefore, the parameters constituting the CNN and each transformer block included in the in-loop filter can be trained in advance and then used in the same manner as stored in the image encoding device and the image decoding device). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to provide the technique of Kang to the modified system of MA and Chen to provide a video coding method and device that applies a current video block to an attention module of a deep learning model, a transformer, and uses a transformer-based in-loop filter accordingly, in order to improve video encoding efficiency and enhance video quality[Kang; para 0006].
Regarding claim 2, MA discloses wherein the U-Net architecture comprises:
one or more pixel shuffle operations; and one or more concatenators for concatenating tensors in a depth domain([see Fig. 1]- The pixel shuffle operations are considered to be implicit in the up-sampling blocks "Up" depicted in Fig. 1; The concatenators are disclosed in the "Base network" in Fig. 1 ).
Regarding claim 3, MA discloses wherein the one or more concatenators are configured to extract features from an output of a last residual block during a down-sampling to an input of a first residual block during up-sampling([see Fig. 1]- The pixel shuffle operations are considered to be implicit in the up-sampling blocks "Up" depicted in Fig. 1; The concatenators are disclosed in the "Base network" in Fig. 1 ).
Regarding claim 4, MA discloses a first stage comprising: a depth-wise 1x1 convolution applied across input channels; a non-linearity operation; and a 3x3 group convolution; and a second stage comprising a feed-forward network, the feed-forward network having a number of hidden layers that is larger than a number of input channels to the feed-forward network([see in Fig. 2]- the channel number of the
residual block is added to 96, and the 1 x1 convolution layer is added before
every 3x3 convolution layer) ; Chen discloses a second stage comprising a feed-forward network, the feed-forward network having a number of hidden layers([see in Fig. 7]-hidden layers)
Regarding claim 5, MA discloses wherein the number of input channels to the feedforward network is larger than 2([see in Fig. 2]- the channel number of the
residual block is added to 96, and the 1 x1 convolution layer is added before
every 3x3 convolution layer).
Regarding claim 6, MA discloses wherein the number of hidden layers is constrained to be equal to a number divisible by at least one of 8 or 16([see in Fig. 2]- the channel number of the residual block is added to 96, and the 1 x1 convolution layer is added before every 3x3 convolution layer).
Regarding claim 7, MA discloses the U-Net architecture further comprising a skip connection, the skip connection being associated with a residual block, a first stage of the residual block, or a second stage of the residual block)[see in section-1]- The architecture has 12 residual blocks in total. The kernel of the convolution layer in residual block is 3x3, and the channel number of the convolution layer in residual block is 64).
Regarding claim 8, Chen discloses wherein the U-Net architecture comprises a plurality of layers of spatial decomposition, and wherein a number of the plurality of layers is equal to an integer number within a range from 2 to log2 of a minimum of a height of the current block and a width of the current block, inclusive([see in Fig. 1-2 and 7]- FIG. 7 is a block diagram illustrating a Fully Connected Neural Network (FC-NN), in accordance with some implementations of the present disclosure. A simple FC- NN, as shown in FIG. 7, is consisting of input layer, output layer, and multiple hidden layers).
Regarding claim 9, Chen discloses wherein the U-Net architecture comprises a neural network encoder and a neural network decoder, and wherein a stride value of a
convolution in a down-sampling path of the neural network encoder is equal to a stride
value of a respective pixel shuffle of the neural network decoder and is an integer value([para 00317;00154]- the original pictures/videos in the selected data set(s) may be up- sampled or down-sampled to different resolutions before encoded and reconstructed).
Regarding claim 10, Chen discloses wherein the U-Net architecture comprises a neural network encoder and a neural network decoder, and wherein a number of layers and a down-sampling factor of the neural network encoder is equal to a number of layers of the neural network decoder and an up-sampling factor, respectively, of the neural network decoder([para 00317;00154]- the original pictures/videos in the selected data set(s) may be up- sampled or down-sampled to different resolutions before encoded and reconstructed; see also para[00137]).
Regarding claim 11, Chen discloses wherein the U-Net architecture comprises a neural network encoder and a neural network decoder, and wherein a number of layers and a down-sampling factor of the neural network encoder is different than a number of layers of the neural network decoder and an up-sampling factor, respectively, of the neural network decoder([para 00317;00154]- the original pictures/videos in the selected data set(s) may be up- sampled or down-sampled to different resolutions before encoded and reconstructed; see also para[00137])..
Regarding claim 12, MA discloses wherein a layer of the U-Net architecture comprises one or more residual blocks followed by one or more transform blocks([see in Fig. 1]-transform block).
Regarding claim 13, Chen discloses wherein the layer of the U-Net architecture comprises two residual blocks followed by one transform block([para 0091-0094]-transform block and residual block).
Regarding claim 14, Chen discloses wherein the U-Net architecture comprises a plurality of decomposition layers and each of a subset of the plurality of decomposition layers comprises one or more transform blocks, the subset of the plurality of decomposition layers comprising a fewer number of decomposition layers than the plurality of decomposition layers([para 0091-0094]-transform block and residual block).
Regarding claim 14, Chen discloses wherein the U-Net architecture comprises a neural network encoder and a neural network decoder, and wherein the neural network encoder comprises a first number of residual blocks and a first number of transform blocks and the neural network decoder comprises a second number of residual blocks and a second number of transform blocks, and wherein at least one of the first number of residual blocks and the second number of residual blocks or the first number of transform blocks and the second number of transform blocks are different([para 0091-0094]-transform block and residual block).
Regarding claim 15, Chen discloses wherein the U-Net architecture comprises a neural network encoder and a neural network decoder, and wherein the neural network encoder comprises a first number of residual blocks and a first number of transform blocks and the neural network decoder comprises a second number of residual blocks and a second number of transform blocks, and wherein at least one of the first number of residual blocks and the second number of residual blocks or the first number of transform blocks and the second number of transform blocks are different([para 0091-0094]-transform block and residual block).
Regarding claim 16, Chen discloses wherein a first layer of decomposition of the U-Net architecture comprises a first number of residual blocks and a first number of transform blocks and a second layer of decomposition of the U-Net architecture comprises a second number of residual blocks and a second number of transform blocks, and wherein at least one of the first number of residual blocks and the second number of residual blocks or the first number of transform blocks and the second number of transform blocks are different([see in Fig. 4C and para 0091-0094]- as illustrated in FIG. 4C, video encoder 20 may use quad-tree partitioning to decompose the luma, Cb, and Cr residual blocks of a CU into one or more luma, Cb, and Cr transform blocks. A transform block is a rectangular (square or non-square) block of samples on which the same transform is applied. A transform unit (TU) of a CU may comprise a transform block of luma samples, two corresponding transform blocks of chroma samples, and syntax elements used to transform the transform block samples).
Regarding claim 17, MA discloses wherein a number of multi-scale processing layers of the U-Net architecture is variable ([see in Fig. 1, abstract and see in Section 2.2]- a convolutional neural network-based in-loop filtering method with QP based models. In this contribution, we modify the training strategy and Unet-Based model).
Regarding claim 18, Chen discloses further comprising training the neural network-based in-loop filter([abstract]- reconstructing, from a video bitstream, a picture frame that includes a luma component, a first and a second chroma components, and applying a trained neural network based in-loop filter to the reconstructed picture frame).
Regarding claim 19, the claim is interpreted and rejected for the same reason as set forth in claim 1. Hence; all limitations for claim 19 have been met in claim 1.
Regarding claim 20, the claim is interpreted and rejected for the same reason as set forth in claim 1. Hence; all limitations for claim 20 have been met in claim 1.
Citation of Pertinent Prior Art
The prior art are made of record and not relied upon but considered pertinent to applicant’s disclosure:
1. XU et al., US 2022/0116639 A1, discloses pruning methods and apparatuses for neural network based video coding.
2. YANG et. al., US 2023/0412806 A1, discloses method, apparatus, and computer program product quantizing neural networks.
3. Li et al., US 2023/0051066 A1, discloses a method implemented by a video coding apparatus. The method includes applying a neural network (NN) filter to an unfiltered sample of a video unit to generate a filtered sample.
4. KIM Mun Churl , US 2019/0230354 A1, discloses a decoding method and apparatus including convolutional neural network (CNN)-based in-loop filter.
Conclusion
THIS ACTION IS MADE FINAL. 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.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MD NAZMUL HAQUE whose telephone number is (571)272-5328. The examiner can normally be reached IFW.
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, David Czekaj can be reached at 5712727327. 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.
/MD N HAQUE/Primary Examiner, Art Unit 2487