DETAILED ACTION
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 .
Applicant’s response to the Non-final Office Action dated 12/18/2025, filed with the office on 02/23/2026, has been entered and made of record.
Information Disclosure Statement
The information disclosure statement (“IDS”) filed on 03/02/2026 have been reviewed and the listed references have been considered.
Response to Amendment
In light of Applicant’s amendments of the claims, the 35 U.S.C. 112(a) rejections of record for failing to comply with the written description requirement have been withdrawn.
Status of Claims
Claims 1-4, 6 and 8-22 are pending. Claims 1, 11, 15 and 18 are amended. Claims 5 and 7 were previously cancelled.
Response to Arguments
Applicant’s amendment of independent Claims 1, 11, 15 and 18, which has altered the scope of the claims of the instant application, has necessitated the new ground(s) of rejection presented in this office action with respect to claims of the instant application. Accordingly, because Applicant’s arguments are merely directed to the amended portion of the claims, new analyses have been presented below, which make Applicant’s arguments moot.
Consequently, THIS ACTION IS MADE FINAL.
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1, 2, 4, 6, 8-10 and 14-17 are rejected under 35 U.S.C. 103 as being unpatentable over Theis et al. (US 10,623,775 B1), in view of Chaudhuri et al. (US 2019/0045168 A1), in further view of Rossato et al. (US 2013/0322530 A1) and still in further view of Chin et al. (US 2020/0074293 A1).
Regarding claim 1, Theis teaches, A method for modifying an input image, wherein the method is applied (Theis, col 2, lines 9-11: “methods to compress image or video data using lossy image compression algorithms”) to a computer device and comprises: (Theis, col 13, lines 40-41: “the application of computer vision techniques to image processing or video encoding”) (Theis, col 12, lines 51-55: “the neural network is functioning as an optimization process, by limiting the number of connection between neurons and/or layers thus enabling a neural network approach to work with high dimensional data such as images”) at least one stage including image down-sampling and filtering of the down-sampled image; (Theis, col 7, lines 48-51: “video data is convolved and spatially downsampled while at the same time increasing the number of channels to m. This is followed by additional convolution(s) with m filters”) at least one stage of image up-sampling, (Theis, col 8, lines 66-67: “a reorganization of coefficients resulting in an upsampling of the video data”) wherein the image down-sampling is performed by applying a strided convolution; (Theis, col 2, lines 65-66: “The compressed video data can be downsampled using a stride in a convolution”). However, Theis does not explicitly teach, generating an output image by processing the input image with a neural network, wherein the output image is a correction image, and including a plurality of trainable weights, wherein the plurality of trainable weights determine a contribution of overlapping sub-regions of the input image to a sample of the down-sampled image, and wherein filtering and downsampling are jointly and simultaneously performed on the input image based on the plurality of trainable weights; modifying the input image by combining the input image with the correction image to obtain an enhanced image, wherein the correction image comprises less degrees of freedom than the enhanced image; and wherein an activation function of the neural network is a leaky rectified linear unit activation function.
In an analogous field of endeavor, Chaudhuri teaches, generating an output image by processing the input image with a neural network, (Chaudhuri, ¶0022: “at least one convolutional layer, which convolves one or more filters with input feature maps to generate output feature maps… provides for advantageous error correction”) wherein the output image is a correction image, (Chaudhuri, ¶0062: “feature maps are provided to convolution module 527, which performs a convolution, to generate feature image 502”; feature image is interpreted as correction image) and including a plurality of trainable weights, (Chaudhuri, ¶0075: “weights of the network are trained in a training phase implemented at operation 703”) ; modifying the input image by combining the input image with the correction image to obtain an enhanced image, (Chaudhuri, ¶0057: “the output of which (e.g., feature images 502, 505) is added (e.g., via adders 528, 530) with the input (e.g., downscaled intermediate image 217 at the appropriate scaling) to get the final output”) wherein the correction image comprises less degrees of freedom than the enhanced image; (Chaudhuri, ¶0026: “create an image… as video interpolation, panorama generation, virtual reality content creation generation with 6 degrees of freedom (DoF)”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis using the teachings of Chaudhuri to introduce modifying an image by adding neural network generated correction image. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of automatically enhancing an input image. Therefore, it would have been obvious to combine the analogous arts Theis and Chaudhuri to obtain the above-described limitations of claim 1. However, the combination of Theis and Chaudhuri does not explicitly teach, wherein the plurality of trainable weights determine a contribution of overlapping sub-regions of the input image to a sample of the down-sampled image, and wherein filtering and downsampling are jointly and simultaneously performed on the input image based on the plurality of trainable weights and wherein an activation function of the neural network is a leaky rectified linear unit activation function.
In another analogous field of endeavor, Rossato teaches, wherein the plurality of trainable weights (Rossato, ¶0179: “weights are calculated by the encoder in order to reduce an amount of entropy of residual data”) determine a contribution of overlapping sub-regions of the input image to a sample of the down-sampled image, (Rossato, ¶0146: “a bilinear downsampling filter with a scale factor of 2 along both dimensions, and weight parameters a and b just depend on the relative time durations of the spans of the higher level of quality that are downsampled into the span of the lower level of quality”) and wherein filtering and downsampling are jointly and simultaneously performed on the input image based on the plurality of trainable weights (Rossato, ¶0173: “the encoder leverages a tweaked filtering method, selectively modifying ("tweaking") the results of a linear downsampling operation and optimizing tweaks based on the resulting entropy of residual data”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis in view of Chaudhuri using the teachings of Rossato to introduce a downsampling based on a trainable weight. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of optimized downsampling of the input image. Therefore, it would have been obvious to combine the analogous arts Theis, Chaudhuri and Rossato to obtain the above-described limitations in claim 1. However, the combination of Theis, Chaudhuri and Rossato does not explicitly teach, wherein an activation function of the neural network is a leaky rectified linear unit activation function.
In yet another analogous field of endeavor, Chin teaches, wherein an activation function of the neural network is a leaky rectified linear unit activation function. (Chin, ¶0017: “Rectified Linear Unit (ReLU) and leaky ReLU activation functions”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis in view of Chaudhuri in further view of Rossato using the teachings of Chin to introduce a leaky ReLu activation function. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of better gradient flow and faster convergence. Therefore, it would have been obvious to combine the analogous arts Theis, Chaudhuri, Rossato and Chin to obtain the invention in claim 1.
Regarding claim 2, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, The method according to claim 1, wherein the strided convolution has a stride of 2. (Theis, col 10, line 40: “convolutions with 128 filters and a stride of 2”).
Regarding claim 4, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, The method according to claim 1, wherein the neural network is parametrized according to a value of a parameter indicative of an amount or type of distortion of the input image. (Theis, col 4, lines 44-48: “All three components may have parameters used to optimize a tradeoff between using a small number of bits (e.g., high compression and low bandwidth) for an encoded frame of video or an encoded image and having small distortion”).
Regarding claim 6, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, The method according to claim 1, wherein the image down-sampling is performed by applying padded convolution. (Theis, col 2, lines 65-66: “(Theis, col 7, lines 47-49: “The normalized and mirror padded video data is convolved and spatially downsampled”).
Regarding claim 8, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, The method according to claim 1, wherein the correction image and the input image have the same vertical and horizontal dimensions, (Chaudhuri, ¶0024: “intermediate image and the feature image (which are at the same resolution) are combined”) and the correction image is a difference image (Chaudhuri, ¶0057: “feature image indicates an image having image difference information”) and the combining the input image with the correction image is performed by addition (Chaudhuri, ¶0057: “feature images 502, 505) is added (e.g., via adders 528, 530) with the input (e.g., downscaled intermediate image 217”) of the difference image to the input image. (Chaudhuri, ¶0057: “feature image indicates an image having image difference information such that a feature image may not generate a fully formed image”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin using the additional teachings of Chaudhuri to introduce a difference image. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of adding the difference image to compensate for image distortion in the input image. Therefore, it would have been obvious to combine the analogous arts Theis, Chaudhuri, Rossato and Chin to obtain the invention of claim 8.
Regarding claim 9, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, A method for reconstructing (Theis, col 17, lines 4-5: “A loop filter 745 can then be applied to the reconstructed block to reduce distortion”) an encoded image from a bitstream, the method including: decoding the encoded image from the bitstream, (Theis, col. 4, 46-49: “optimize a tradeoff between using a small number of bits (e.g., high compression and low bandwidth) for an encoded frame of video or an encoded image and having small distortion when the encoded frame of video or encoded image is decoded”) and applying the method for modifying the input image according to claim 1 with the input image being the decoded image. (Theis, col 16, lines 63-67: “The reverse encoder 735 may be configured to decode the output of the encoder 110… the output of the reverse encoder 735 may be used as input”).
Regarding claim 10, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, A method for reconstructing a compressed image of a video, comprising: reconstructing an image using an image prediction (Theis, col 16, lines 57-60: “In intra-frame prediction, a sample is predicted from reconstructed pixels within the same frame for the purpose of reducing the residual error that is coded by the encoder 110”) based on a reference image (Theis, col 16, lines 60-62: “Inter-frame prediction relates to predicting the pixel values in a block of a picture relative to data of a previously coded picture”) stored in a memory, (Theis, col 11, lines 34-35: “at least one memory 410 may be configured to store data”) applying the method for modifying the input image according to claim 1 with the input image being the reconstructed image, (Theis, col 16, lines 33-35: “video encoder 625 can be used to encode input video data 5. As shown in FIG. 7A, dashed lines represent a reconstruction path”) and storing the modified input image into the memory as a reference image. (Theis, col 4 lines 62-64: “output of the encoder (e.g., compressed video data 10) is the code used to represent an image and is stored”).
Regarding claim 14, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, A non-transitory computer-readable medium comprising computer programs which are executed by one or more processors and cause the one or more processors to perform the method according to claim 1. (Theis, col. 2, lines 45-49:” a non-transitory computer readable medium includes code segments that when executed by a processor cause the processor to perform steps”).
Regarding claim 15, it recites a device with elements corresponding to the steps of the method recited in claim 1. Therefore, the recited elements of device claim 15 are mapped to the proposed combination in the same manner as the corresponding steps in method claim 1. Additionally, the rationale and motivation to combine Theis, Chaudhuri, Rossato and Chin presented in rejection of claim 1, apply to this claim. Theis further teaches, A device for modifying an input image, (Theis, col. 2, lines 20-22: “one device configured to decode the compressed video data using an inverse algorithm based on the lossy compression algorithm”) comprising a memory coupled to a processor and having computer-executable instructions stored thereon; and the processor configured to execute the instructions (Theis, col. 11, lines 24-27: “The at least one processor 405 may be utilized to execute instructions stored on the at least one memory 410, so as to thereby implement the various features and functions described herein”).
Regarding claim 16, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, A device for reconstructing an encoded image from a bitstream, comprising: a decoder configured to decode the encoded image from the bitstream, to obtain a decoded image (Theis, col. 4, 46-49: “optimize a tradeoff between using a small number of bits (e.g., high compression and low bandwidth) for an encoded frame of video or an encoded image and having small distortion when the encoded frame of video or encoded image is decoded”) and the device configured to modify the decoded image according to claim 15. (Theis, col 16, lines 63-67: “The reverse encoder 735 may be configured to decode the output of the encoder 110… the output of the reverse encoder 735 may be used as input”).
Regarding claim 17, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, A device for reconstructing a compressed image of a video, comprising: an adder configured to reconstruct an image using an image prediction (Theis, col 16, lines 57-60: “In intra-frame prediction, a sample is predicted from reconstructed pixels within the same frame for the purpose of reducing the residual error that is coded by the encoder 110”) based on a reference image (Theis, col 16, lines 60-62: “Inter-frame prediction relates to predicting the pixel values in a block of a picture relative to data of a previously coded picture”) stored in a memory, (Theis, col 11, lines 34-35: “at least one memory 410 may be configured to store data”) the device configured to modify the decoded image according to claim 16, and a memory storing the modified image as a reference image. (Theis, col 4 lines 62-64: “output of the encoder (e.g., compressed video data 10) is the code used to represent an image and is stored”).
Claim 3 is rejected under 35 U.S.C. 103 as being unpatentable over Theis et al. (US 10,623,775 B1), in view of Chaudhuri et al. (US 2019/0045168 A1), in further view of Rossato et al. (US 2013/0322530 A1), yet in further view of Chin et al. (US 2020/0074293 A1) and still in further view of Jansson et al. (US 2020/0042879 A1).
Regarding claim 3, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, The method according to claim 1. However, The combination of Theis, Chaudhuri, Rossato and Chin does not explicitly teach, wherein the neural network is based on a U-net, and wherein for establishing the neural network, the U-net is modified by introducing a skip connection to the U-net, the skip connection is adapted to connect the input image with the output image.
In an analogous field of endeavor, Jansson teaches, wherein the neural network is based on a U-net, (Jansson, ¶0011: “the neural network system includes a U-Net”) and wherein for establishing the neural network, the U-net is modified by introducing a skip connection to the U-net, the skip connection is adapted to connect the input image with the output image. (Jansson, ¶0054: “the U-Net architecture herein adds additional skip connections between layers at the same hierarchical level in the encoder and decoder. This enables low-level information to flow directly from the high-resolution input to the high-resolution output.”)
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis in view of Chaudhuri, in further view of Rossato and yet in further view of Chin using the teachings of Jansson to introduce a U-net with skip connection. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of flow of information from the input to the output. Therefore, it would have been obvious to combine the analogous arts Theis, Chaudhuri, Rossato, Chin and Jansson to obtain the invention of claim 3.
Claims 11 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Theis et al. (US 10,623,775 B1), in view of Chaudhuri et al. (US 2019/0045168 A1), in further view of Rossato et al. (US 2013/0322530 A1), yet in further view of Chin et al. (US 2020/0074293 A1) and still in further view of Riedmueller et al. (US 2017/0131119 A1).
Regarding claim 11, Theis teaches, A method for training a neural network for modifying a distorted image, (Theis, col 12, lines 9-16: “Neural networks are then trained until performance improvement… for a fixed rate-distortion trade-off”) wherein the method is applied to a computer device and comprises: (Theis, col 13, lines 40-41: “the application of computer vision techniques to image processing or video encoding”)(Theis, col 12, lines 51-55: “the neural network is functioning as an optimization process, by limiting the number of connection between neurons and/or layers thus enabling a neural network approach to work with high dimensional data such as images”) at least one stage including an image down-sampling and a filtering of the down-sampled image; (Theis, col 7, lines 48-51: “video data is convolved and spatially downsampled while at the same time increasing the number of channels to m. This is followed by additional convolution(s) with m filters”) at least one stage of an image up-sampling, (Theis, col 8, lines 66-67: “a reorganization of coefficients resulting in an upsampling of the video data”) wherein the image down-sampling is performed by applying a strided convolution; (Theis, col 2, lines 65-66: “The compressed video data can be downsampled using a stride in a convolution”). However, Theis does not explicitly teach, inputting, to the neural network, pairs of a distorted image as a target input and a target output image which is based on an original image, wherein the target output image is a correction image, and including a plurality of trainable weights, wherein the plurality of trainable weights determine a contribution of overlapping sub-regions of an input image to a sample of the down-sampled image, and wherein filtering and downsampling are jointly and simultaneously performed on the input image based on the plurality of trainable weights; adapting at least one parameter of the filtering based on the inputted pairs; modifying the input image by computing the input image by combining the input image with the correction image to obtain an enhanced image, wherein the correction image comprises less degrees of freedom than the enhanced image; and wherein an activation function of the neural network is a leaky rectified linear unit activation function, and wherein according to the leaky rectified linear unit activation function.
In an analogous field of endeavor, Chaudhuri teaches, inputting, to the neural network, pairs of a distorted image as a target input and a target output image (Chaudhuri, ¶0057: “feature images 502, 505) is added (e.g., via adders 528, 530) with the input (e.g., downscaled intermediate image 217”) which is based on an original image, wherein the target output image is a correction image, (Chaudhuri, ¶0062: “feature maps are provided to convolution module 527, which performs a convolution, to generate feature image 502”; feature image is interpreted as correction image) and including a plurality of trainable weights, (Chaudhuri, ¶0075: “weights of the network are trained in a training phase implemented at operation 703”) (Chaudhuri, ¶0057: “the output of which (e.g., feature images 502, 505) is added (e.g., via adders 528, 530) with the input (e.g., downscaled intermediate image 217 at the appropriate scaling) to get the final output”) wherein the correction image comprises less degrees of freedom than the enhanced image; (Chaudhuri, ¶0026: “create an image… as video interpolation, panorama generation, virtual reality content creation generation with 6 degrees of freedom (DoF)”; also see the analysis on response to arguments section).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis using the teachings of Chaudhuri to introduce modifying an image by adding neural network generated correction image. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of automatically enhancing an input image. Therefore, it would have been obvious to combine the analogous arts Theis and Chaudhuri to obtain the invention of claim 11. However, The combination of Theis and Chaudhuri does not explicitly teach, wherein an activation function of the neural network is a leaky rectified linear unit activation function and wherein the plurality of trainable weights determine a contribution of overlapping sub-regions of an input image to a sample of the down-sampled image, and wherein filtering and downsampling are jointly and simultaneously performed on the input image based on the plurality of trainable weights; adapting at least one parameter of the filtering based on the inputted pairs.
In another analogous field of endeavor, Rossato teaches, wherein the plurality of trainable weights (Rossato, ¶0179: “weights are calculated by the encoder in order to reduce an amount of entropy of residual data”) determine a contribution of overlapping sub-regions of an input image to a sample of the down-sampled image, (Rossato, ¶0146: “a bilinear downsampling filter with a scale factor of 2 along both dimensions, and weight parameters a and b just depend on the relative time durations of the spans of the higher level of quality that are downsampled into the span of the lower level of quality”) and wherein filtering and downsampling are jointly and simultaneously performed on the input image based on the plurality of trainable weights (Rossato, ¶0173: “the encoder leverages a tweaked filtering method, selectively modifying ("tweaking") the results of a linear downsampling operation and optimizing tweaks based on the resulting entropy of residual data”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis in view of Chaudhuri using the teachings of Rossato to introduce a downsampling based on a trainable weight. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of optimized downsampling of the input image. Therefore, it would have been obvious to combine the analogous arts Theis, Chaudhuri and Rossato to obtain the above-described limitations in claim 1. However, the combination of Theis, Chaudhuri and Rossato does not explicitly teach, wherein an activation function of the neural network is a leaky rectified linear unit activation function and adapting at least one parameter of the filtering based on the inputted pairs.
In another analogous field of endeavor, Chin teaches, wherein an activation function of the neural network is a leaky rectified linear unit activation function. (Chin, ¶0017: “Rectified Linear Unit (ReLU) and leaky ReLU activation functions”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis in view of Chaudhuri and in further view of Rossato using the teachings of Chin to introduce a leaky ReLu activation function. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of better gradient flow and faster convergence. Therefore, it would have been obvious to combine the analogous arts Theis, Chaudhuri, Rossato and Chin to obtain the above-described limitations in claim 11. However, the combination of Theis, Chaudhuri, Rossato and Chin does not explicitly teach adapting at least one parameter of the filtering based on the inputted pairs.
In yet another analogous field of endeavor, Riedmueller teaches, adapting at least one parameter of the filtering based on the inputted pairs. (Riedmueller, ¶0110: “The loop filter 16 may be implemented as adaptive loop filter. At least one parameter, filter parameter or coefficient of the loop filter 16 may be changed over time to adapt to changing signal characteristics”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis in view of Chaudhuri in further view of Rossato and yet in further view of Chin using the teachings of Riedmueller to introduce adaptive filtering. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of controlling a parameter of a filter for optimized filtering. Therefore, it would have been obvious to combine the analogous arts Theis, Chaudhuri, Rossato, Chin and Riedmueller to obtain the invention of claim 11.
Regarding claim 18, it recites a device with elements corresponding to the steps of the method recited in claim 11. Therefore, the recited elements of device claim 18 are mapped to the proposed combination in the same manner as the corresponding steps in method claim 11. Additionally, the rationale and motivation to combine Theis, Chaudhuri, Rossato, Chin and Riedmueller presented in rejection of claim 11, apply to this claim.
Claims 12, 13, 19 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Theis et al. (US 10,623,775 B1), in view of Chaudhuri et al. (US 2019/0045168 A1), in further view of Rossato et al. (US 2013/0322530 A1), yet in further view of Chin et al. (US 2020/0074293 A1), still in further view of Riedmueller et al. (US 2017/0131119 A1) and even in further view of Nguyen et al. (US 2020/0075148 A).
Regarding claim 12, Theis in view of Chaudhuri in further view of Rossato, yet in further view of Chin and still in further view of Riedmueller teaches, The method according to claim 11. However, the combination of Theis, Chaudhuri, Rossato, Chin and Riedmueller does not explicitly teach, wherein the adapting of the at least one parameter of the filtering is based on a loss function corresponding to Mean Squared Error (MSE).
In an analogous field of endeavor, Nguyen teaches, wherein the adapting of the at least one parameter of the filtering is based on a loss function corresponding to Mean Squared Error (MSE). (Nguyen, ¶0215: “The optimized dose, that was generated using the dose influence array and Chambolle-Pock algorithm, is used to minimize against the predicted dose distribution with a mean squared error loss”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis in view of Chaudhuri, in further view of Rossato, yet in further view of Chin and still in further view of Riedmueller using the teachings of Nguyen to introduce a loss function. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of quantifying the model’s performance. Therefore, it would have been obvious to combine the analogous arts Theis, Chaudhuri, Rossato, Chin, Riedmueller and Nguyen to obtain the invention of claim 12.
Regarding claim 13, Theis in view of Chaudhuri in further view of Rossato, yet in further view of Chin and still in further view of Riedmueller teaches, The method according to claim 11, wherein the adapting of the at least one parameter of the filtering is based (Theis, col. 11, lines 61-64: “iterates over channels of the coefficients for a single image z. GSMs can be useful building blocks for modeling filter responses of natural images”). However, the combination of Theis, Chaudhuri, Rossato, Chin and Riedmueller does not explicitly teach, on a loss function including a weighted average of squared errors.
In an analogous field of endeavor, Nguyen teaches, on a loss function including a weighted average of squared errors (Nguyen, ¶0182: “a multi-objective optimization problem and mathematically can be expressed as the multi-objective weighted least squares function”).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of the claimed invention to modify Theis in view of Chaudhuri, in further view of Rossato, yet in further view of Chin and still in further view of Riedmueller using the teachings of Nguyen to introduce weighted average of squared errors. A person skilled in the art would be motivated to combine the known elements as described above and achieve the predictable result of quantifying the model’s performance. Therefore, it would have been obvious to combine the analogous arts Theis, Chaudhuri, Rossato, Chin, Riedmueller and Nguyen to obtain the invention of claim 13.
Regarding claim 19, it recites a device with elements corresponding to the steps of the method recited in claim 12. Therefore, the recited elements of device claim 19 are mapped to the proposed combination in the same manner as the corresponding steps in method claim 12. Additionally, the rationale and motivation to combine Theis, Chaudhuri, Rossato, Chin, Riedmueller and Nguyen presented in rejection of claim 12, apply to this claim.
Regarding claim 20, it recites a device with elements corresponding to the steps of the method recited in claim 13. Therefore, the recited elements of device claim 20 are mapped to the proposed combination in the same manner as the corresponding steps in method claim 13. Additionally, the rationale and motivation to combine Theis, Chaudhuri, Rossato, Chin, Riedmueller and Nguyen presented in rejection of claim 13, apply to this claim.
Claims 21 and 22 are rejected under 35 U.S.C. 103 as being unpatentable over Theis et al. (US 10,623,775 B1), in view of Chaudhuri et al. (US 2019/0045168 A1), in further view of Rossato et al. (US 2013/0322530 A1), yet in further view of Chin et al. (US 2020/0074293 A1) and still in further view of Nguyen et al. (US 2020/0075148 A).
Regarding claim 21, Theis in view of Chaudhuri, in further view of Rossato and in further view of Chin teaches, The method according to claim 1. However, the combination of Theis, Chaudhuri, Rossato and Chin does not explicitly teach, wherein the activation function of the neural network is a softplus activation function.
In an analogous field of endeavor, Nguyen teaches, wherein the activation function of the neural network is a softplus activation function. (Nguyen, ¶0212: “The final activation layer as the softplus activation, as the output data is non-negative”).
Regarding claim 22, it recites a device with elements corresponding to the steps of the method recited in claim 21. Therefore, the recited elements of device claim 22 are mapped to the proposed combination in the same manner as the corresponding steps in method claim 21. Additionally, the rationale and motivation to combine Theis, Chaudhuri, Rossato, Chin and Nguyen presented in rejection of claim 21, apply to this claim.
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 MEHRAZUL ISLAM whose telephone number is (571)270-0489. The examiner can normally be reached Monday-Friday: 8am-5pm.
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, Saini Amandeep can be reached at (571) 272-3382. 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.
/MEHRAZUL ISLAM/Examiner, Art Unit 2662
/AMANDEEP SAINI/Supervisory Patent Examiner, Art Unit 2662
Prosecution Insights