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 .
Claims 1-20 are presented for examination.
Drawings
The drawings are objected to as failing to comply with 37 CFR 1.84(p)(5) because they do not include the following reference sign(s) mentioned in the description: “multi-rate image compression encoder (1611)” in paragraph [0125]. Corrected drawing sheets in compliance with 37 CFR 1.121(d) are required in reply to the Office action to avoid abandonment of the application. Any amended replacement drawing sheet should include all of the figures appearing on the immediate prior version of the sheet, even if only one figure is being amended. Each drawing sheet submitted after the filing date of an application must be labeled in the top margin as either “Replacement Sheet” or “New Sheet” pursuant to 37 CFR 1.121(d). If the changes are not accepted by the examiner, the applicant will be notified and informed of any required corrective action in the next Office action. The objection to the drawings will not be held in abeyance.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP § 2146 et seq. for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The filing of a terminal disclaimer by itself is not a complete reply to a nonstatutory double patenting (NSDP) rejection. A complete reply requires that the terminal disclaimer be accompanied by a reply requesting reconsideration of the prior Office action. Even where the NSDP rejection is provisional the reply must be complete. See MPEP § 804, subsection I.B.1. For a reply to a non-final Office action, see 37 CFR 1.111(a). For a reply to final Office action, see 37 CFR 1.113(c). A request for reconsideration while not provided for in 37 CFR 1.113(c) may be filed after final for consideration. See MPEP §§ 706.07(e) and 714.13.
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The actual filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/apply/applying-online/eterminal-disclaimer.
Claims 1-20 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1-9 of U.S. Patent No. 12,335,487. Although the claims at issue are not identical, they are not patentably distinct from each other because the claims of the present invention are similar in scope to the claims of the U.S. Patent No. 12,335,487. For example, the table below shows the similarities and difference between the instant application and the patent.
(Instant Application) 18/122,645
US 12,335,487 B2
Claim 1) A method for image processing, comprising:
Claim 1) A method for image processing, comprising:
determining, from a coded bitstream that carries a compressed image as an input to a Compression Domain Computer Vision Task Framework (CDCVFT), a value of a tunable hyperparameter indicating a compression rate of the compressed image, the compressed image being generated by a neural network based encoder according to the value of the tunable hyperparameter;
decoding, from a coded bitstream that carries a compressed image, an index that points to a value in a set of values of a parameter, a value change of the parameter adjusting a compression rate of the compressed image, the compressed image being generated by a neural network based encoder based on the parameter,
Claim 8) The method of claim 1, wherein the parameter is a hyperparameter.
Claim 9) The method of claim 1, wherein the tunable hyperparameter is for weighting a distortion in a calculation of a rate distortion loss.
the parameter weighting a distortion and a rate in a calculation of a rate-distortion (R-D) loss;
inputting the value of the tunable hyperparameter to a multi-rate compression domain computer vision task decoder, the multi-rate compression domain computer vision task decoder comprising one or more neural networks for performing a computer vision task from compressed images according to corresponding values of the tunable hyperparameter that are used for generating the compressed images; and
inputting the value of the parameter to a multi-rate compression domain computer vision task decoder in a compression domain computer vision task framework (CDCVTF), the multi-rate compression domain computer vision task decoder comprising one or more neural networks for performing a computer vision task from compressed images according to corresponding values of the parameter that are used for generating the compressed images at multiple different compression rates; and
generating, by the multi-rate compression domain computer vision task decoder, a computer vision task result according to the compressed image in the coded bitstream and the value of the tunable hyperparameter.
generating, by the multi-rate compression domain computer vision task decoder, a computer vision task result according to the compressed image in the coded bitstream which is compressed at a corresponding compression rate from the multiple different compression rates, the compression rate being based on the value of the parameter,
Claim 2) The method of Claim 1, wherein the generating the computer vision task result comprises:
wherein the generating the computer vision task result includes:
converting, by a first neural network in the multi-rate compression domain computer vision task decoder, the value of the tunable hyperparameter to a tensor; inputting the tensor to one or more layers in a second neural network in the multi-rate compression domain computer vision task decoder; and generating, by the second neural network, the computer vision task result according to the compressed image and the tensor.
converting, by a first neural network in the multi-rate compression domain computer vision task decoder, the value of the parameter to a tensor; inputting the tensor to one or more layers in a second neural network in the multi-rate compression domain computer vision task decoder; and generating, by the second neural network, the computer vision task result according to the compressed image and the tensor.
Claim 3) The method of claim 2, wherein the first neural network comprises one or more convolution layers.
Claim 2) The method of claim 1, wherein the first neural network comprises one or more convolution layers.
Claim 4) The method of claim 2, wherein the first neural network comprises a convolution layer with an activation function.
Claim 3) The method of claim 1, wherein the first neural network comprises a convolution layer with an activation function.
Claim 5) The method of claim 2, wherein the second neural network is configured to generate the computer vision task result without generating a reconstructed image from the compressed image.
Claim 4) The method of claim 1, wherein the second neural network is configured to generate the computer vision task result without generating a reconstructed image from the compressed image.
Claim 6) The method of claim 2, wherein the second neural network is configured to generate a reconstructed image from the compressed image, and generate the computer vision task result from the reconstructed image.
Claim 5) The method of claim 1, wherein the second neural network is configured to generate a reconstructed image from the compressed image, and generate the computer vision task result from the reconstructed image.
Claim 7) The method of claim 1, wherein the neural network based encoder is based on an encoder model in a neural image compression (NIC) framework, and the multi-rate compression domain computer vision task decoder is based on a decoder model in the NIC framework, the NIC framework is trained end-to-end.
Claim 6) The method of claim 1, wherein the neural network based encoder is based on an encoder model in a neural image compression (NIC) framework, and the multi-rate compression domain computer vision task decoder is based on a decoder model in the NIC framework, the NIC framework is trained end-to-end.
Claim 8) The method of claim 1, wherein a decoder model of the multi-rate compression domain computer vision task decoder is trained separate from an encoder model of the neural network based encoder.
Claim 7) The method of claim 1, wherein a decoder model of the multi-rate compression domain computer vision task decoder is trained separate from an encoder model of the neural network based encoder.
Claim 10) The method of claim 1, wherein the computer vision task comprises at least one of image classification, image denoising, object detection and super resolution.
Claim 9) The method of claim 1, wherein the computer vision task comprises at least one of image classification, image denoising, object detection and super resolution.
Claim 11) An apparatus for image processing, comprising processing circuitry configured to:
Claim 1) A method for image processing, comprising:
determine, from a coded bitstream that carries a compressed image as an input to a compression domain computer vision task framework (CDCVFT), a value of a tunable hyperparameter indicating a compression rate of the compressed image, the compressed image being generated by a neural network based encoder according to the value of the tunable hyperparameter;
decoding, from a coded bitstream that carries a compressed image, an index that points to a value in a set of values of a parameter, a value change of the parameter adjusting a compression rate of the compressed image, the compressed image being generated by a neural network based encoder based on the parameter,
Claim 8) The method of claim 1, wherein the parameter is a hyperparameter.
Claim 19) The apparatus of claim 11, wherein the tunable hyperparameter is for weighting a distortion in a calculation of a rate distortion loss.
the parameter weighting a distortion and a rate in a calculation of a rate-distortion (R-D) loss;
input the value of the tunable hyperparameter to a multi-rate compression domain computer vision task decoder, the multi-rate compression domain computer vision task decoder comprising one or more neural networks for performing a computer vision task from compressed images according to corresponding values of the tunable hyperparameter that are used for generating the compressed images; and
inputting the value of the parameter to a multi-rate compression domain computer vision task decoder in a compression domain computer vision task framework (CDCVTF), the multi-rate compression domain computer vision task decoder comprising one or more neural networks for performing a computer vision task from compressed images according to corresponding values of the parameter that are used for generating the compressed images at multiple different compression rates; and
generate, by the multi-rate compression domain computer vision task decoder, a computer vision task result according to the compressed image in the coded bitstream and the value of the tunable hyperparameter.
generating, by the multi-rate compression domain computer vision task decoder, a computer vision task result according to the compressed image in the coded bitstream which is compressed at a corresponding compression rate from the multiple different compression rates, the compression rate being based on the value of the parameter,
Claim 12) The apparatus of claim 11, wherein the processing circuitry is configured to:
wherein the generating the computer vision task result includes:
convert, by a first neural network in the multi-rate compression domain computer vision task decoder, the value of the tunable hyperparameter to a tensor; input the tensor to one or more layers in a second neural network in the multi-rate compression domain computer vision task decoder; and generate, by the second neural network, the computer vision task result according to the compressed image and the tensor.
converting, by a first neural network in the multi-rate compression domain computer vision task decoder, the value of the parameter to a tensor; inputting the tensor to one or more layers in a second neural network in the multi-rate compression domain computer vision task decoder; and generating, by the second neural network, the computer vision task result according to the compressed image and the tensor.
Claim 13) The apparatus of claim 12, wherein the first neural network comprises one or more convolution layers.
Claim 2) The method of claim 1, wherein the first neural network comprises one or more convolution layers.
Claim 14) The apparatus of claim 12, wherein the first neural network comprises a convolution layer with an activation function.
Claim 3) The method of claim 1, wherein the first neural network comprises a convolution layer with an activation function.
Claim 15) The apparatus of claim 12, wherein the second neural network is configured to generate the computer vision task result without generating a reconstructed image from the compressed image.
Claim 4) The method of claim 1, wherein the second neural network is configured to generate the computer vision task result without generating a reconstructed image from the compressed image.
Claim 16) The apparatus of claim 12, wherein the second neural network is configured to generate a reconstructed image from the compressed image, and generate the computer vision task result from the reconstructed image.
Claim 5) The method of claim 1, wherein the second neural network is configured to generate a reconstructed image from the compressed image, and generate the computer vision task result from the reconstructed image.
Claim 17) The apparatus of claim 11, wherein the neural network based encoder is based on an encoder model in a neural image compression (NIC) framework, and the multi-rate compression domain computer vision task decoder is based on a decoder model in the NIC framework, the NIC framework is trained end-to-end.
Claim 6) The method of claim 1, wherein the neural network based encoder is based on an encoder model in a neural image compression (NIC) framework, and the multi-rate compression domain computer vision task decoder is based on a decoder model in the NIC framework, the NIC framework is trained end-to-end.
Claim 18) The apparatus of claim 11, wherein a decoder model of the multi-rate compression domain computer vision task decoder is trained separate from an encoder model of the neural network based encoder.
Claim 7) The method of claim 1, wherein a decoder model of the multi-rate compression domain computer vision task decoder is trained separate from an encoder model of the neural network based encoder.
Claim 20) The apparatus of claim 11, wherein the computer vision task comprises at least one of image classification, image denoising, object detection and super resolution.
Claim 9) The method of claim 1, wherein the computer vision task comprises at least one of image classification, image denoising, object detection and super resolution.
The differences between the instant application and the U.S. Patent No. 12,335,487 amount to either a change in wording that is ultimately directed to the same limitation (e.g., “a value of a tunable hyperparameter indicating a compression rate” vs. “a value change of the parameter adjusting a compression rate” “at multiple different compression rates” “wherein the parameter is a hyperparameter”), a descriptor being used in a different location in the claims but otherwise are describing the same thing (e.g., “a compressed image as an input to a computer domain computer vision task framework (CDCVFT)” vs. “inputting the value of the parameter to a… decoder in a compression domain computer vision task framework (CDCVTF)”), the moving of limitations between the independent and dependent claims (e.g., limitations of Claim 2 of instant application integrated into Claim 1 of patent, limitations of Claim 9 in instant application located in independent claim 1 of patent and limitations of Claim 8 in patent located in independent claim 1 of instant application), or the addition of a layer of abstraction regarding the hyperparameter in the patent (i.e., determine a value of a hyperparameter from the coded bitstream vs. determine an index pointing to a value of a hyperparameter from the coded bitstream), none of which expand the scope of the instant application to be patentably distinct from the U.S. Patent No. 12,335,487.
Claim Rejections - 35 USC § 103
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 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 1-6, 9-16, and 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over Choi et al. (“Variable Rate Deep Image Compression with a Conditional Autoencoder”, published 10/30/2019), hereinafter Choi, in view of Torfason et al. (“Towards Image Understanding from Deep Compression without Decoding”, published 03/16/2018), hereinafter Torfason.
Regarding Claim 1, Choi teaches A method for image processing, comprising:
determining, from a coded bitstream that carries a compressed image as an input to a Compression Domain Computer Vision Task Framework (CDCVFT), a value of a tunable hyperparameter indicating a compression rate of the compressed image, the compressed image being generated by a neural network based encoder according to the value of the tunable hyperparameter (Choi: “We decode the compressed bitstream. We also retrieve λ and ∆ used in encoding from the compressed bitstream. We restore the quantized latent representation from the decoded integer values by multiplying them with the quantization bin size ∆.” [Section 3.3. Inference]; “We provide two knobs to vary the rate. First, we employ a conditional autoencoder, conditioned on the Lagrange multiplier λ that adapts the rate, and optimize the rate-distortion Lagrangian for various λ values in one conditional model.” [Fig. 1]; “To avoid training and deploying multiple networks, we propose training one conditional autoencoder, conditioned on the Lagrange multiplier λ. The network takes λ as a conditioning input parameter, along with the input image, and produces a compressed image with varying rate and distortion depending on the conditioning value of λ.” [Section 3.1. Conditional autoencoder]); and
inputting the value of the tunable hyperparameter to a multi-rate compression domain computer vision task decoder, the multi-rate compression domain computer vision task decoder comprising one or more neural networks for decompressing images from compressed images according to corresponding values of the tunable hyperparameter that are used for generating the compressed images (Choi: “The restored latent representation is then fed to the decoder to reconstruct the image. The value of λ used in encoding is again used in all deconvolutional layers, for conditional generation.” [Section 3.3. Inference]; “To avoid training and deploying multiple networks, we propose training one conditional autoencoder, conditioned on the Lagrange multiplier λ. The network takes λ as a conditioning input parameter, along with the input image, and produces a compressed image with varying rate and distortion depending on the conditioning value of λ.” [Section 3.1. Conditional autoencoder]).
However, Choi fails to expressly disclose one or more neural networks for performing a computer vision task from compressed images according to corresponding values of the tunable hyperparameter that are used for generating the compressed images; and generating, by the multi-rate compression domain computer vision task decoder, a computer vision task result according to the compressed image in the coded bitstream and the value of the tunable hyperparameter.
In the same field of endeavor, Torfason teaches one or more neural networks for performing a computer vision task from compressed images according to corresponding values of the tunable hyperparameter that are used for generating the compressed images (Torfason: “Motivated by recent work on deep neural network (DNN)-based image compression methods showing potential improvements in image quality, savings in storage, and bandwidth reduction, we propose to perform image understanding tasks such as classification and segmentation directly on the compressed representations produced by these compression methods.” [Abstract]; “We note here that the encoder of the convolutional autoencoder produces a compressed representation (feature map) of dimensions w/8 × h/8 × C, where w and h are the spatial dimensions of the input image, and the number of channels C is a hyperparameter related to the rate R.” [Section 3.1 Deep Compression Architecture]); and
generating, by the multi-rate compression domain computer vision task decoder, a computer vision task result according to the compressed image in the coded bitstream and the value of the tunable hyperparameter (Torfason: “Motivated by recent work on deep neural network (DNN)-based image compression methods showing potential improvements in image quality, savings in storage, and bandwidth reduction, we propose to perform image understanding tasks such as classification and segmentation directly on the compressed representations produced by these compression methods. Since the encoders and decoders in DNN-based compression methods are neural networks with feature-maps as internal representations of the images, we directly integrate these with architectures for image understanding.” [Abstract]; “We note here that the encoder of the convolutional autoencoder produces a compressed representation (feature map) of dimensions w/8 × h/8 × C, where w and h are the spatial dimensions of the input image, and the number of channels C is a hyperparameter related to the rate R.” [Section 3.1 Deep Compression Architecture]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have incorporated one or more neural networks for performing a computer vision task from compressed images according to corresponding values of the tunable hyperparameter that are used for generating the compressed images; and generating, by the multi-rate compression domain computer vision task decoder, a computer vision task result according to the compressed image in the coded bitstream and the value of the tunable hyperparameter, as taught by Torfason to the method of Choi because both of these methods are directed towards deep neural network-based image compression. In making this combination and employing the neural network decoder to perform computer vision tasks from compressed images according to a compression rate hyperparameter, it would allow the method of Choi to “bypass[[es]] decoding of the compressed representation into RGB space and reduces computational cost” (Torfason: [Abstract]).
Regarding Claim 2, Choi and Torfason teach the method of Claim 1, wherein the generating the computer vision task result comprises:
converting, by a first neural network in the multi-rate compression domain computer vision task decoder, the value of the tunable hyperparameter to a tensor (Choi: “After selecting λ ∈ Λ, we do one-hot encoding of λ and use it in all conditional convolutional layers to encode a latent representation of the input. Then, we perform regular deterministic quantization on the encoded representation with the selected quantization bin size ∆. The quantized latent representation is then finally encoded into a compressed bitstream with entropy coding, e.g., arithmetic coding; we additionally need to store the values of the conditioning variables, λ and ∆, used in encoding.” [Section 3.3. Inference]; “To implement a conditional autoencoder, we develop the conditional convolution, conditioned on the Lagrange multiplier λ, as shown in Figure 4. Let Xi be a 2-dimensional (2-D) input feature map of channel i and Yj be a 2-D output feature map of channel j. Let Wi,j be a 2-D convolutional kernel for input channel i and output channel j.” [Section 3.1. Conditional autoencoder]);
inputting the tensor to one or more layers in a second neural network in the multi-rate compression domain computer vision task decoder (Choi: “We decode the compressed bitstream. We also retrieve λ and ∆ used in encoding from the com pressed bitstream. We restore the quantized latent representation from the decoded integer values by multiplying them with the quantization bin size ∆. The restored latent representation is then fed to the decoder to reconstruct the image. The value of λ used in encoding is again used in all deconvolutional layers, for conditional generation.” [Section 3.3. Inference]); and
generating, by the second neural network, the computer vision task result according to the compressed image and the tensor (Torfason: “we propose to perform image understanding tasks such as classification and segmentation directly on the compressed representations produced by these compression methods. Since the encoders and decoders in DNN-based compression methods are neural networks with feature-maps as internal representations of the images, we directly integrate these with architectures for image understanding.” [Abstract]).
Regarding Claim 3, Choi and Torfason teach the method of Claim 2, wherein the first neural network comprises one or more convolution layers (Choi: “After selecting λ ∈ Λ, we do one-hot encoding of λ and use it in all conditional convolutional layers to encode a latent representation of the input.” [Section 3.3. Inference]).
Regarding Claim 4, Choi and Torfason teach the method of Claim 2, wherein the first neural network comprises a convolution layer with an activation function (Choi: See [Fig. 7], in which convolutional layers in the encoder are followed by a leaky ReLU layer).
Regarding Claim 5, Choi and Torfason teach the method of Claim 2, wherein the second neural network is configured to generate the computer vision task result without generating a reconstructed image from the compressed image (Torfason: “In this paper, we explore another promising advantage of learned compression algorithms compared to engineered ones, namely the amenability of the compressed representation they produce to learning and inference without reconstruction (see Fig. 2). Specifically, instead of recon structing an RGB image from the (quantized) compressed representation and feeding it to a network for inference (e.g., classification or segmentation), one uses a modified network that bypasses reconstruction of the RGB image.” [Section 1. Introduction]).
Regarding Claim 6, Choi and Torfason teach the method of Claim 2, wherein the second neural network is configured to generate a reconstructed image from the compressed image, and generate the computer vision task result from the reconstructed image (Torfason: “Results are shown for ResNet-50 (where reconstructed/decoded RGB images are used as input) and for cResNet-51 and cResNet-39 (where compressed representations are used as input).” [Fig. 3]).
Regarding Claim 9, Choi and Torfason teach the method of Claim 1, wherein the tunable hyperparameter is for weighting a distortion in a calculation of a rate distortion loss (Choi: “The network takes λ as a conditioning input parameter, along with the input image, and produces a compressed image with varying rate and distortion depending on the conditioning value of λ. To this end, the rate and distortion terms in (4) and (5) are altered into [Equation] for λ ∈ Λ, where Λ is a pre-defined finite set of Lagrange multiplier values, and then we minimize the following combined objective function: [Equation 7].” [Section 3.1. Conditional autoencoder]).
Regarding Claim 10, Choi and Torfason teach the method of Claim 1, wherein the computer vision task comprises at least one of image classification, image denoising, object detection and super resolution (Torfason: “Motivated by recent work on deep neural network (DNN)-based image compression methods showing potential improvements in image quality, savings in storage, and bandwidth reduction, we propose to perform image understanding tasks such as classification and segmentation directly on the compressed representations produced by these compression methods.” [Abstract]).
Regarding Claims 11-16 and 19-20, they are apparatus claims that correspond with the method of Claims 1-6 and 9-10. Therefore, they are rejected for the same reasons as Claims 1-6 and 9-10 above.
Claims 7-8 and 17-18 are rejected under 35 U.S.C. 103 as being unpatentable over Choi in view of Torfason, in further view of Besenbruch et al. (US 20230154055 A1, provisional application filed 07/20/2020).
Regarding Claim 7, Choi and Torfason teach the method of Claim 1, wherein the neural network based encoder is based on an encoder model in a neural image compression (NIC) framework, and the multi-rate compression domain computer vision task decoder is based on a decoder model in the NIC framework (Choi: “we train and deploy only one variable-rate image compression network implemented with a conditional autoencoder.” [Abstract]; See [Fig. 1]).
However, they fail to expressly disclose wherein the NIC framework is trained end-to-end.
In the same field of endeavor, Besenbruch teaches wherein the NIC framework is trained end-to-end (Besenbruch: “The method may be one wherein the segmentation algorithm neural network is trained end-to-end with the first neural network and the second neural network.” [0235]; “According to a fifth aspect of the invention, there is provided a computer implemented method of training a first neural network and a second neural network, the neural networks being for use in lossy image or video compression, transmission and decoding, the method including the steps of: (i) receiving an input training image; (ii) encoding the input training image using the first neural network, to produce a latent representation; (iii) quantizing the latent representation to produce a quantized latent; (iv) using the second neural network to produce an output image from the quantized latent, wherein the output image is an approximation of the input image” [0087]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have incorporated wherein the NIC framework is trained end-to-end, as taught by Besenbruch to the method of Choi and Torfason because both methods are directed towards neural network-based image compression. End-to-end training is a common training technique in deep learning. In making this combination and training the framework end-to-end, it would allow the method of Choi and Torfason to reduce training time (Besenbruch: [0808]).
Regarding Claim 8, Choi and Torfason teach the method of Claim 1. However, they fail to expressly disclose wherein a decoder model of the multi-rate compression domain computer vision task decoder is trained separate from an encoder model of the neural network based encoder.
In the same field of endeavor, Besenbruch teaches wherein a decoder model of the multi-rate compression domain computer vision task decoder is trained separate from an encoder model of the neural network based encoder (Besenbruch: “The method may be one wherein the segmentation algorithm neural network is trained separately to the first neural network and to the second neural network.” [0234]; “According to a fifth aspect of the invention, there is provided a computer implemented method of training a first neural network and a second neural network, the neural networks being for use in lossy image or video compression, transmission and decoding, the method including the steps of: (i) receiving an input training image; (ii) encoding the input training image using the first neural network, to produce a latent representation; (iii) quantizing the latent representation to produce a quantized latent; (iv) using the second neural network to produce an output image from the quantized latent, wherein the output image is an approximation of the input image” [0087]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have incorporated wherein a decoder model of the multi-rate compression domain computer vision task decoder is trained separate from an encoder model of the neural network based encoder, as taught by Besenbruch to the method of Choi and Torfason because both methods are directed towards neural network-based image compression. Independent component training is a common training technique in deep learning. In making this combination and training the encoder and decoder separately, it would allow the method of Choi and Torfason to prevent gradients from impacting the other components when training for compression and computer vision tasks (Besenbruch: [0578]).
Regarding Claims 17 and 18, they are apparatus claims that correspond with the method of Claims 7 and 8. Therefore, they are rejected for the same reasons as Claims 7 and 8 above.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Yang et al. (“Variable Rate Deep Image Compression with Modulated Autoencoder”) teaches adaptive deep image compression autoencoding to address a specific R-D tradeoff when utilizing multiple rates of compression.
Toderici et al. (“Full Resolution Image Compression with Recurrent Neural Networks”) teaches a set of full-resolution lossy image compression methods based on neural networks with variable compression rates without requiring retraining.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MEGAN E HWANG whose telephone number is (703)756-1377. The examiner can normally be reached Monday-Thursday 10:00AM-7:30PM ET.
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, Jennifer Welch can be reached at (571) 272-7212. 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.
/M.E.H./Examiner, Art Unit 2143
/JENNIFER N WELCH/Supervisory Patent Examiner, Art Unit 2143
Prosecution Insights