CTNF 18/933,406 CTNF 85908 DETAILED ACTION Allowable Subject Matter 12-151-08 AIA 07-43 12-51-08 Claim s 2-3 and 13-14 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims. Claim Rejections - 35 USC § 102 07-07-aia AIA 07-07 The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action: A person shall be entitled to a patent unless – 07-08-aia AIA (a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention. 07-15 AIA Claim s 1, 6, 8 and 12, are rejected under 35 U.S.C. 102( a)(1 ) as being anticipated by NOROUZZADEH et al. (Pub. No. US 2024/0104339 hereinafter “NOR”) . Regarding claims 1 and 12, NOR teaches a method for controlling an electronic apparatus for performing a computer vision task (classification) [Para. 46], the method comprising: receiving a corrupted image from a camera [Para. 46 “ The system may include an input image 301 that is received from one or more sensors, such as a camera, radar, Lidar, etc . The hyper model 305 may identify the type of input corruption, e.g., motion blur or Gaussian noise .” And “ The main classifier model 309 may perform classification on the input image ”]; identifying a corruption type (type of corruption) of the corrupted image using a corruption identification module (hyper model 305) [Para. 46 “ The hyper model 305 may identify the type of input corruption, e.g., motion blur or Gaussian noise. In one example, a simple classifier model that takes the Fourier transform of the input image as input and predicts the type of corruption in the image or if it is a clean image. ”]; obtaining normalisation parameters (BN statistics) associated with the identified corruption type from a codebook (lookup table) [Para. 46 “ The system may include a BN statistics lookup table (T) 307 that contains corresponding values for each type of corruption .”; “ Once the hyper model detects the corruption, it can dynamically choose (or generate) corresponding BN statistics (or, in a more general procedure, even parameters or architecture of the main model) from the lookup table in real-time to make the main model insensitive to that type of corruption ”]; updating a computer vision model (pre-trained network), trained to perform the task (classification), by replacing normalisation parameters of the computer vision model with the obtained normalisation parameters (BN statistics) [Para. 24 “ Given that corruption type can be easily detected using the Fourier domain, the system and method may adopt the BN statistic update method such that it can change the BN values dynamically based on the current corruption ”; “ Based on the detected corruption, the corresponding BN statistics are taken from the BN stat lookup table and the pre-trained network BNs are updated accordingly ” and Para. 46 “ The main classifier model 309 may perform classification on the input image ”]; and performing the computer vision task (classification) using the updated computer vision model [Para. 24 “ Finally, the original image is fed to the dynamically updated pre-trained network ”; Para. 46 “ Upon updating the model of the classifier 309 , the classifier 309 may output the corresponding classification 311 .”]. Regarding claim 6, NOR teaches wherein the computer vision model is a neural network model (DNN), and wherein the normalisation parameters include at least one of batch normalisation, BatchNorm, parameters or layer normalisation, LayerNorm, parameters [Para. 26 and 46]. Regarding claim 8, NOR teaches wherein the computer vision task is a computer vision task selected from a list of computer vision tasks including object detection, object recognition, semantic segmentation [Para. 26 and 30] . Claim Rejections - 35 USC § 103 07-20-aia AIA 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. 07-21-aia AIA Claim s 4, 5, and 15 are rejected under 35 U.S.C. 103 as being unpatentable over NOROUZZADEH et al. (Pub. No. US 2024/0104339 hereinafter “NOR”) in view of Drolia et al. (Pub. No. US 2022/0114747) . Regarding claims 4 and 5, NOR teaches wherein the corruption identification module (hyper model) includes a machine learning model (DNN) trained to estimate a corruption type (type of corruption) using a corrupted image [Para. 26 and 46]. However, NOR doesn’t explicitly teach the rest of claim limitations. Drolia teaches wherein identifying the corruption type of the corrupted image using a corruption identification module comprises: inputting the corrupted image (input image) to the machine learning model (artificial neural network) to estimate the corruption type (distortion type) from the corrupted image (input image) [Para. 31, 39 and 32]. It would have been obvious to one of ordinary skill in the art before the effective filing date to modify NOR’s corruption identification module (hyper model) by incorporating Drolia’s teaching of inputting the corrupted image (input image) to a machine learning model that outputs a corruption type, instead of feeding only a Fourier transform of the image to the corruption detector. This modification improves NOR by enabling direct learned corruption classification from camera images, thereby allowing the same BN-statics lookup process to operate when spatial image feature provide useful degradation information . 07-21-aia AIA Claim s 7 and 10 rejected under 35 U.S.C. 103 as being unpatentable over NOROUZZADEH et al. (Pub. No. US 2024/0104339 hereinafter “NOR”) in view of LIM et al. (Pub. NO. US 2024/0119360) . Regarding claim 7, NOR teaches DNN trained network [Para. 26]. However, NOR doesn’t explicitly teach the neural network model is a convolutional neural network model. LIM teaches the neural network model is a convolutional neural network model [para. 35]. It would have been obvious to one of ordinary skill in the art before the effective filing date to modify NOR’s DNN trained on the original vision task by implementing Lim’s teaching of a convolutional neural network model with batch-normalization processing for image-based inference. This medication improves NOR by using convolutional visual feature extraction for object detection/classification inputs, thereby improving suitability for camera image computer vision tasks. Regarding claim 10, NOR doesn’t explicitly teach the claim limitations. However, LIM teaches wherein the normalisation layers include at least one of batch normalisation, BatchNorm, layers and layer normalisation, LayerNorm, layers [Para. 39]. It would have been obvious to one of ordinary skill in the art before the effective filing date to modify NOR’s BN statistics lookup table and model update process by incorporating Lim’s teaching of normalization layers that use those statistics, thereby enabling predictable layer level adaption of the task model . 07-21-aia AIA Claim s 9 and 11 are rejected under 35 U.S.C. 103 as being unpatentable over NOROUZZADEH et al. (Pub. No. US 2024/0104339 hereinafter “NOR”) in view of Segu et al. (Pub. No. US 2023/0122207) . Regarding claim 9, NOR teaches generating the codebook by: providing a pre-trained computer vision model (pre-trained model) and a training data set, wherein the training data set (train data) comprises, for each corruption type of a plurality of corruption types, a plurality of corrupted images and corresponding labels associated with the computer vision task the model has been trained to perform, re-training the pre-trained computer vision model, for each corruption type; generating the codebook to associate each recognizable corruption type to the corresponding normalisation layers parameters (BN statistics) [Para. 26, 6, and 46]. However, NOR doesn’t explicitly teach the rest of claim limitations. Segu teaches using the plurality of corrupted images and corresponding labels (object categories) by updating only normalisation layers of the pre-trained computer vision model, extracting the normalisation layers of the re-trained computer vision model for each corruption type [Para. 33, 22, and 23]. It would have been obvious to one of ordinary skill in the art before the effective filing date to modify NOR to incorporate convolutional neural network, feature as taught by Segu; because the modification reduces retraining cost while preserving the task-model weights for corrupted image computer vision. Regarding claim 11, NOR teaches generating the codebook (lookup table) by: providing a pre-trained computer vision model, corrupted images, and corresponding corruption type labels estimated by the corruption identification module, updating normalisation layers of the pre-trained computer vision model based on the corrupted images and the corresponding corruption type labels, extracting the updated normalisation layers for each estimated corruption type, and generating a codebook to associate each recognizable corruption type to the corresponding normalisation layers parameters [Para. 46, fig. 2 and related description]. However, NOR doesn’t explicitly teach updating normalisation layers using a test-time adaptation algorithm [Para. 23, 54, and 57]. It would have been obvious to one of ordinary skill in the art before the effective filing date to modify NOR’s BN-statistics lookup table and inference time update process by applying Segu’s test time adaptation algorithm to update the normalization layer parameters for test data grouped by NOR’s estimated corruption type, then storing those updated BN statistics in NOR’s lookup table for the corresponding type of corruption. This modification improves NOR by making the codebook entries target data adapted at inference time, thereby improving corrupted image robustness when the incoming corruption distribution changes. Conclusion Any inquiry concerning this communication or earlier communications from the examiner should be directed to SOLOMON G BEZUAYEHU whose telephone number is (571)270-7452. The examiner can normally be reached on Monday-Friday 10 AM-7 PM. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, O’Neal Mistry can be reached on 313-446-4912 . The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300. Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-0101 (IN USA OR CANADA) or 571-272-1000. /SOLOMON G BEZUAYEHU/ Primary Examiner, Art Unit 2666 Application/Control Number: 18/933,406 Page 2 Art Unit: 2674 Application/Control Number: 18/933,406 Page 3 Art Unit: 2674 Application/Control Number: 18/933,406 Page 4 Art Unit: 2674 Application/Control Number: 18/933,406 Page 5 Art Unit: 2674 Application/Control Number: 18/933,406 Page 6 Art Unit: 2674 Application/Control Number: 18/933,406 Page 7 Art Unit: 2674 Application/Control Number: 18/933,406 Page 8 Art Unit: 2674 Application/Control Number: 18/933,406 Page 9 Art Unit: 2674
Prosecution Insights