DETAILED ACTION
This office action dated 07/29/2026 replaces the prior action issued 07/13/2026. The current office action issued 07/29/2026 supersedes any previous action issued.
The United States Patent & Trademark Office appreciates the response filed for the
current application that is submitted on 06/05/2026. The United States Patent & Trademark
Office reviewed the following documents submitted and has made the following comments below.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 06/05/2026 has been entered.
Priority
Acknowledgment is made that this application is a CON of application no. PCT/JP2022/021329 filed on 05/25/2022, which further claim for domestic priority under 35 U.S.C.119 (e) based on the provisional applications PRO 63/193,785 filed on 05/27/2021.
Information Disclosure Statement
The IDS dated 11/17/2023 has been considered and placed in the application file.
Overview
Claims 1-7 are pending in this application and have been considered below.
Claims 1-3 and 5-7 are rejected.
Claim 4 is objected to.
Applicant Arguments:
In regards to the argument on Argument 1, Applicant/s state/s “ Accordingly, Tong merely teaches simply adding noise to a source image and teaches that its impact can become locally asymmetric. However, Tong does not teach locally adding noise to a source image, and as such, the Examiner's position that paragraph [0077] of Tong teaches "adding noise to the source image locally" is clearly incorrect. Therefore, it is respectfully submitted that Tong does not disclose or suggest "generating a first image by adding noise to a first area that is a part of an original image," as required by the above-noted features of claim 1.” (See Remarks Pg 7 ¶05-¶06). Therefore the 35 U.S.C 103 rejection on the amended claims should be withdrawn.
In regards to the argument on Argument 2, Applicant/s state/s “ Applicant respectfully submits that, based on the plain meaning of the terms "labels" and "images," one of ordinary skill in the art would recognize that weighted addition of labels differs from weighted addition of images." (See Remarks Pg 8 ¶02). Therefore the 35 U.S.C 103 rejection on the amended claims should be withdrawn.
In regards to the argument on Argument 3, Applicant/s state/s “ Applicant respectfully submits that the cited references neither indicate the use of the second ratio between sizes, nor do they indicate that the second ratio between sizes is used to perform weighted addition of labels.” (See Remarks Pg 08 ¶04). Therefore the 35 U.S.C 103 rejection on the pending claims should be withdrawn.
In regards to the argument on Argument 4, Applicant/s state/s “ it is it is respectfully submitted that any combination of Tong and Ando fails to teach "generating a first training label for the first image by weighted addition of a first base label corresponding to a correct label of the original image and a second base label corresponding to an incorrect label of the original image at a second ratio that is a ratio between a size of the first area and a size of the second area," "generating a second training label for the second image by weighted addition of the first base label and the second base label at an inverse ratio of the second ratio," and "generating a combined training label for the combined image by weighted addition of the first training label and the second training label at the first ratio," as required by the above-noted features of claim 1.” (See Remarks Pg 08 ¶05). Therefore the 35 U.S.C 103 rejection on the amended claims and its dependent claims should be withdrawn.
Examiners Responses:
In response to Argument 1, see remarks filed 06/05/2026, Applicant’s arguments, see Remarks, filed 06/05/2026, with respect to the U.S.C 103 rejections of Claim 1 have been considered but are moot in view of new ground(s) of rejection caused by the amendments. A new ground(s) of rejection is made for claims 1-3 and 5-7, are rejected under 35 U.S.C. 103 as unpatentable over Tong et al (US Patent Pub 2020/0074234 A1 hereafter referred to as Tong) in view of Ando (US Patent Pub 2016/0026900 A1hereafter referred to as Ando) in further view of Zhong (Zhong, Zhun, et al. "Random erasing data augmentation." Proceedings of the AAAI conference on artificial intelligence. Vol. 34. No. 07. 2020).
The Examiner finds that Tong teaches on the claim language “an original image” and Ando teaches “a first area” in the amended independent claims 1 and 6 .
The Examiner interprets under broadest reasonable interpretation that the Claim states “generating a first image by adding noise to a first area that is part of an original image”, specifically, “the first area” within the claim can be interpreted as adding noise to the source image locally as taught in ¶0077 by Tong. Furthermore, the claim does not limit what “the first area” is; thus allowing a person skilled in the art to interpret broadly that the first area in which noise is added to the image can be interpreted as adding noise to the source image locally. The examiner finds that Tong in combination with Ando does disclose adding noise to a first area in an original image. Tong does disclose generating a training image by adding noise to the original image in ¶0075-¶0078, and ¶0161. Ando discloses the original image being split to focus on different areas in Fig 3 and ¶0128. Therefore, the Examiner interprets that the combination of Tong and Ando teaches the main concept of adding noise to an area of an original image, the additional details of the functions of the main concepts as stated above by the applicant in the amendments is taught by Zhong in the details of the rejection below. The Examiner will maintain prior art Tong and Ando and details of the rejection are below.
In response to Argument 2, see remarks filed 06/05/2026, the Examiner respectfully disagrees. In response to that Argument that based on the plain meaning of the terms "labels" and "images," one of ordinary skill in the art would recognize that weighted addition of labels differs from weighted addition of images, the Examiner respectfully disagrees. No special definition of weighted addition is found in the present specification, and, absent a special definition, Examiner is obligated to take the broadest reasonable interpretation not in conflict with the specification. It is noted that the feature upon which applicant relies (i.e., “weighted addition of two labels”) has been given its broadest reasonable interpretation. MPEP 2111-2111.01. Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993). The examiner finds that Tong does teach a first and second training label added to the weighted images in ¶0086 as Asource and Anoise being the labels being applied to the images, and that the first and second images are weighted before they are combined in ¶0015 and ¶0086 in Tong. Under broadest reasonable interpretation the examiner interprets that one of ordinary skill in the art would find that, “weighted addition of the labels,” can be interpreted as two images with labels being weighted and combined as taught by Tong. The specification is silent as to the weighted addition of the labels and the timing of the addition; the specification does not prohibit such an interpretation; therefore, Examiner's interpretation is both reasonable and not in conflict with the specification, and the limitation is met by the prior art. Therefore, the Examiner will maintain prior art Tong and Ando and details of the rejection are below.
In response to Argument 3, see remarks filed 06/05/2026, The Examiner interprets under broadest reasonable interpretation that the Claim states “original image at a second ratio”, specifically, “a second ratio” within the claim can be interpreted as the source image having varying noise ratios that could include 0.75, 0.5, 0.4, 0.3, 0.2 or others in ¶0007 by Tong. Furthermore, the claim does not limit what value “a second ratio” is; thus allowing a person skilled in the art to interpret broadly that an original image at a second ratio can be interpreted as source image having varying noise ratios that could include 0.75, 0.5, 0.4, 0.3, 0.2. The Examiner interprets under broadest reasonable interpretation that the Claim states “a ratio between a size of the first area and a size of the second area”, specifically, “a ratio between a size” within the claim can be interpreted as the source image having a large segmentation area size and an second area that does not include the content of the first area being small in ¶0056 and ¶0159 by Ando. Furthermore, the claim does not limit what value “a ratio between size” is; thus allowing a person skilled in the art to interpret broadly that a ratio between size can be interpreted as the first area being large and the second area being small. Under broadest reasonable interpretation of the claims the examiner interprets that Tong does teach a first and second training label added to the weighted images in ¶0086 as Asource and Anoise being the labels being applied to the images, and that the first and second images are weighted before they are combined in ¶0015 and ¶0086 in Tong. Under broadest reasonable interpretation the examiner interprets that one of ordinary skill in the art would find that, “weighted addition of the labels,” can be interpreted as two images with labels being weighted and combined as taught by Tong. The specification is silent as to the weighted addition of the labels and the timing of the addition; the specification does not prohibit such an interpretation; therefore, Examiner's interpretation is both reasonable and not in conflict with the specification, and the limitation is met by the prior art. The Examiner interprets under broadest reasonable interpretation that the Claim states “first base label and the second base label at an inverse ratio in of the second ratio”, specifically, “inverse ratio in of the second ratio” within the claim can be interpreted as using an inverse function to determine the pixel noise within the images in ¶0018 by Tong. Furthermore, the claim does not limit the value of what “inverse ratio in of the second ratio” is; thus allowing a person skilled in the art to interpret broadly that an inverse ratio in of the second ratio is using an inverse function to determine the pixel noise within the images. Tong discloses the original image at a second ratio in ¶0074 while Ando teaches a ratio between a size of the first area and a size of the second area in ¶0056 and ¶0159. Tong then continues to disclose generating a second training label in ¶0032-¶0033 for the second image by weighted addition in ¶0015 and ¶0086 and of the first base label and the second base label at an inverse ratio in ¶0018 of the second ratio in ¶0074. Therefore Tong in combination with Ando does teach the limitations of Claim 1 pertaining to the second ratio between sizes and the addition of weighted labels. Therefore, the Examiner will maintain prior art Tong and Ando and details of the rejection are below.
In response to Argument 4, see remarks filed 06/05/2026, Applicant’s arguments, see Remarks, filed 06/05/2026, with respect to the U.S.C 103 rejections of Claim 1, Claim 6 and its dependent claims have been considered but are moot in view of new ground(s) of rejection caused by the amendments. A new ground(s) of rejection is made for claims 1-3 and 5-7, are rejected under 35 U.S.C. 103 as unpatentable over Tong et al (US Patent Pub 2020/0074234 A1 hereafter referred to as Tong) in view of Ando (US Patent Pub 2016/0026900 A1hereafter referred to as Ando) in further view of Zhong (Zhong, Zhun, et al. "Random erasing data augmentation." Proceedings of the AAAI conference on artificial intelligence. Vol. 34. No. 07. 2020). See above arguments for explanation and below for the rejection.
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.
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 1-3 and 5-7, are rejected under 35 U.S.C. 103 as unpatentable over Tong et al (US Patent Pub 2020/0074234 A1 hereafter referred to as Tong) in view of Ando (US Patent Pub 2016/0026900 A1hereafter referred to as Ando) in further view of Zhong (Zhong, Zhun, et al. "Random erasing data augmentation." Proceedings of the AAAI conference on artificial intelligence. Vol. 34. No. 07. 2020).
Regarding Claim 1, Tong teaches a training method (Tong ¶0159, ¶0177, ¶0053 discloses a training method) for generating a learning model (Tong Fig 8 and Fig 9, ¶0125, ¶0128 discloses noise trained CNN models) for use in image recognition (Tong ¶0002, discloses a network robust in the recognition of objects in images), the training method comprising:
generating a first image by adding noise (Tong ¶0075-¶0078 discloses generating a noisy training image) an original image (Tong ¶0075-¶0077 discloses adding noise to the original image);
generating a second image by adding noise (Tong ¶0106, ¶0109, ¶0008 discloses generating a second set of noisy training images) in the original image (Tong ¶0075-¶0077 discloses adding noise to the original image);
generating a combined image (Tong ¶0013, ¶0064, disclose generating a combined noise image) by weighted addition (Tong ¶0015, ¶0086 discloses that each image is weighted before they are combined) of the first image (Tong ¶0075-¶0078 discloses generating a noisy training image) and the second image (Tong ¶0106, ¶0109, ¶0008 discloses generating a second set of noisy training images) at a first ratio (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5);
generating a first training label (Tong ¶0032 discloses generating and applying labels for image categories) for the first image by weighted addition (Tong ¶0015, ¶0086 discloses that each image is weighted before they are combined) of a first base label corresponding to a correct label of the original image (Tong ¶0158 discloses the trained network correctly labeling the validation images compared to the untrained network) and a second base label corresponding to an incorrect label (Tong ¶0158 discloses correctly identifying if an image was labeled incorrectly) of the original image at a second ratio (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5, 0.75 etc.);
generating a second training label (Tong ¶0032 discloses generating and applying labels for image categories) for the second image by weighted addition (Tong ¶0015, ¶0086 discloses that each image is weighted before they are combined) of the first base label and the second base label at an inverse ratio (Tong ¶0018 discloses inventing the noise ratio) of the second ratio (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5, 0.75 etc.);
for the combined image (Tong ¶0013, ¶0064, disclose generating a combined noise image) by weighted addition (Tong ¶0015, ¶0086 discloses that each image is weighted before they are combined) of the first training label and the second training label (Tong ¶0032 discloses generating and applying labels for image categories) at the first ratio (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5); and
generating the learning model by machine learning (Tong Fig 8 and Fig 9, ¶0125, ¶0128 discloses noise trained CNN models) using the combined image (Tong ¶0013, ¶0064, disclose generating a combined noise image).
Tong does not explicitly teach to a first area, to a second area that is an area excluding the first area, that is a ratio between a size of the first area and a size of the second area, generating a combined training label, and the combined training label.
Ando is in the same field of generating sets of training pattern. Further, Ando teaches to a first area (Ando Fig 3 discloses segmenting the original area into different areas including IM1), to a second area that is an area excluding the first area (Ando Fig 3 discloses segmenting the original area into different areas including IM1 which is different from IM2), that is a ratio between a size of the first area and a size of the second area (Ando ¶0056, ¶0159 discloses the ratio of the size of the large area in comparison to the objects in the other areas of the original image), generating a combined training label (Ando ¶0135 discloses a label for the combined images used for training), and the combined training label (Ando ¶0135 discloses a label for the combined images used for training).
Therefore, it would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Tong by incorporating different areas of the original image to be processed and using ratios to determine size, as well as combined labeling of the images as taught by Ando, to make an invention that can more robustly identify objects due to partial images being used in the training data; thus, one of ordinary skilled in the art would be motivated to combine the references since an object of the present invention is to increase the amount of training data available without increasing cost (Ando, ¶0008-¶0009).
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Tong and Ando in combination do not explicitly disclose that is part of an original image.
Zhong is in the same field adding noise to datasets for improved robustness. Further, Zhong teaches that is part of (Zhong Fig 1 and Pg 6 Col 2 ¶02-Pg 7 Col 1 ¶01 and Table 3 discloses that the noise is added to area that is part of the original image) an ordinal image.
Therefore, it would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Tong in view of Ando by incorporating noise implementation into a part of the original image as taught by that is part of Zhong to make an invention that can more robustly identify objects due to the inclusion of training data with parts of the original image having noise; thus, one of ordinary skilled in the art would be motivated to combine the references since an object of the present invention is to reduce the risk of over-fitting and make the model robust to occlusion. (Zhong, Abstract).
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Regarding Claim 2, Tong in view of Ando in further view of Zhong teaches the training method according to claim 1, wherein a plurality of combined images (Tong ¶0013, ¶0064, disclose generating a combined noise image) and a plurality (Tong ¶0007, ¶0023-¶0025 discloses a plurality of images used) of combined training labels (Ando ¶0135 discloses a label for the combined images used for training) are generated by generating, for each of a plurality of first areas (Ando ¶0070, ¶0073, ¶0076, ¶0103 discloses segmenting the image into a plurality of areas), the first image (Tong ¶0075-¶0078 discloses generating a noisy training image), the second image (Tong ¶0106, ¶0109, ¶0008 discloses generating a second set of noisy training images), the combined image (Tong ¶0013, ¶0064, disclose generating a combined noise image), the first training label, the second training label, and the combined training label (Ando ¶0135 discloses a label for the combined images used for training), each of the plurality of combined images being the combined image (Tong ¶0013, ¶0064, disclose generating a combined noise image), each of the plurality of combined training labels (Ando ¶0102-¶0104, discloses applying labels to a plurality of images for training data) being the combined training label (Ando ¶0135 discloses a label for the combined images used for training), each of the plurality of first areas (Ando ¶0070-¶0073 discloses a plurality of areas) being the first area (Ando Fig 3 discloses segmenting the original area into different areas including IM1) , and
the learning model (Tong Fig 8 and Fig 9, ¶0125, ¶0128 discloses noise trained CNN models) is generated by machine learning (Tong ¶0006 discloses the formulation of a CNN/DNN) using the plurality of combined images (Tong ¶0013, ¶0064, disclose generating a combined noise image) and the plurality of combined training labels (Ando ¶0135 discloses a label for the combined images used for training). See Claim 1 for rationale, its parent claim.
Regarding Claim 3, Tong in view of Ando in further view of Zhong teaches the training method according to claim 1, wherein
a plurality of combined images (Tong ¶0013, ¶0064, disclose generating a combined noise image) and a plurality of combined training labels (Ando ¶0102-¶0104, discloses applying labels to a plurality of images for training data) are generated by generating the combined image (Tong ¶0013, ¶0064, disclose generating a combined noise image) and the combined training label (Ando ¶0135 discloses a label for the combined images used for training) at each of a plurality of first ratios (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5, 0.75 etc.), each of the plurality (Tong ¶0007, ¶0023-¶0025 discloses a plurality of images used) of combined images being the combined image (Tong ¶0013, ¶0064, disclose generating a combined noise image), each of the plurality of combined training labels (Ando ¶0102-¶0104, discloses applying labels to a plurality of images for training data) being the combined training label(Ando ¶0135 discloses a label for the combined images used for training), each of the plurality of first ratios (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5, 0.75 etc.) being the first ratio (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5), and
the learning model (Tong Fig 8 and Fig 9, ¶0125, ¶0128 discloses noise trained CNN models) is generated by machine learning (Tong ¶0006 discloses the formulation of a CNN/DNN) using the plurality of combined images (Tong ¶0013, ¶0064, disclose generating a combined noise image) and the plurality of combined training labels (Ando ¶0135 discloses a label for the combined images used for training). See Claim 1 for rationale, its parent claim.
Regarding Claim 5, Tong in view of Ando in further view of Zhong teaches the training method according to claim 1, wherein the first ratio (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5) is determined in accordance with a beta distribution of B(a, a) (Tong Fig 4C discloses the beta distribution of the test SSNR and accuracy), where B denotes a beta function (Tong Fig 4C discloses the beta distribution of the test SSNR and accuracy), and a denotes a positive real number (Tong Fig 4C discloses the beta distribution of the test SSNR (0 to 1) and accuracy (0 to 100)). See Claim 1 for rationale, its parent claim.
Regarding Claim 6, Tong teaches a training device (Tong ¶0177 discloses a computing device where the training is preformed) that generates a learning model (Tong Fig 8 and Fig 9, ¶0125, ¶0128 discloses noise trained CNN models) for use in image recognition (Tong ¶0002, discloses a network robust in the recognition of objects in images), the training device comprising:
a processor (Tong ¶0179, ¶0183 discloses a processor); and
memory (Tong ¶0070, ¶179 discloses a memory) wherein
using the memory, the processor (Tong Fig 16 the dashed line discloses the processor and memory bring linked):
generating a first image by adding noise (Tong ¶0075-¶0078 discloses generating a noisy training image) an original image (Tong ¶0075-¶0077 discloses adding noise to the original image);
generating a second image by adding noise (Tong ¶0106, ¶0109, ¶0008 discloses generating a second set of noisy training images) in the original image (Tong ¶0075-¶0077 discloses adding noise to the original image);
generating a combined image (Tong ¶0013, ¶0064, disclose generating a combined noise image) by weighted addition (Tong ¶0015, ¶0086 discloses that each image is weighted before they are combined) of the first image (Tong ¶0075-¶0078 discloses generating a noisy training image) and the second image (Tong ¶0106, ¶0109, ¶0008 discloses generating a second set of noisy training images) at a first ratio (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5);
generating a first training label (Tong ¶0032 discloses generating and applying labels for image categories) for the first image by weighted addition (Tong ¶0015, ¶0086 discloses that each image is weighted before they are combined) of a first base label corresponding to a correct label of the original image (Tong ¶0158 discloses the trained network correctly labeling the validation images compared to the untrained network) and a second base label corresponding to an incorrect label (Tong ¶0158 discloses correctly identifying if an image was labeled incorrectly) of the original image at a second ratio (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5, 0.75 etc.);
generating a second training label (Tong ¶0032 discloses generating and applying labels for image categories) for the second image by weighted addition (Tong ¶0015, ¶0086 discloses that each image is weighted before they are combined) of the first base label and the second base label at an inverse (Tong ¶0018 discloses inventing the noise ratio) ratio of the second ratio (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5, 0.75 etc.);
for the combined image (Tong ¶0013, ¶0064, disclose generating a combined noise image) by weighted addition (Tong ¶0015, ¶0086 discloses that each image is weighted before they are combined) of the first training label and the second training label (Tong ¶0032 discloses generating and applying labels for image categories) at the first ratio (Tong ¶0074 discloses combining the images with varying SSNR values including 0.5); and
generating the learning model by machine learning (Tong Fig 8 and Fig 9, ¶0125, ¶0128 discloses noise trained CNN models) using the combined image (Tong ¶0013, ¶0064, disclose generating a combined noise image).
Tong does not explicitly teach to a first area, to a second area that is an area excluding the first area, that is a ratio between a size of the first area and a size of the second area, generating a combined training label, and the combined training label.
Ando is in the same field of generating sets of training pattern. Further, Ando teaches to a first area (Ando Fig 3 discloses segmenting the original area into different areas including IM1), to a second area that is an area excluding the first area (Ando Fig 3 discloses segmenting the original area into different areas including IM1 which is different from IM2), that is a ratio between a size of the first area and a size of the second area (Ando ¶0056, ¶0159 discloses the ratio of the size of the large area in comparison to the objects in the other areas of the original image), generating a combined training label (Ando ¶0135 discloses a label for the combined images used for training) and the combined training label (Ando ¶0135 discloses a label for the combined images used for training).
Therefore, it would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Tong by incorporating different areas of the original image to be processed and using ratios to determine size, as well as combined labeling of the images as taught by Ando, to make an invention that can more robustly identify objects due to partial images being used in the training data; thus, one of ordinary skilled in the art would be motivated to combine the references since an object of the present invention is to increase the amount of training data available without increasing cost (Ando, ¶0008-¶0009).
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Tong and Ando in combination do not explicitly disclose that is part of an original image.
Zhong is in the same field adding noise to datasets for improved robustness. Further, Zhong teaches that is part of (Zhong Fig 1 and Pg 6 Col 2 ¶02-Pg 7 Col 1 ¶01 and Table 3 discloses that the noise is added to area that is part of the original image) an ordinal image.
Therefore, it would have been obvious to one having ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Tong in view of Ando by incorporating noise implementation into a part of the original image as taught by that is part of Zhong to make an invention that can more robustly identify objects due to the inclusion of training data with parts of the original image having noise; thus, one of ordinary skilled in the art would be motivated to combine the references since an object of the present invention is to reduce the risk of over-fitting and make the model robust to occlusion. (Zhong, Abstract).
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Regarding Claim 7, Tong in view of Ando in view of Zhong teaches a non-transitory computer-readable recording medium (Tong ¶0034, ¶0036 discloses a non-transitory computer readable medium) having recorded thereon a computer program (Tong ¶0182- ¶0183 discloses a computer being programed or loaded with instructions or program code) for causing a computer to execute (Tong ¶0181 discloses a computer device providing instructions to be executed) the training method according to claim 1. See Claim 1 for rationale, its parent claim.
Allowable Subject Matter
Claim 4 is objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form and was overcome including all of the limitations of the base claim and any intervening claims.
Reference Cited
The prior art made of record and not relied upon is considered pertinent to applicant’s
disclosure.
US Patent Pub US-20200342652-A1 to Rowell et al. discloses generating synthetic
image data for machine learning.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RACHEL ROBERTS whose telephone number is (571)272-6413. The examiner can normally be reached Monday- Friday 7:30am- 5:00pm. 32. 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. 33. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Oneal 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. 34. 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.
/RACHEL L ROBERTS/Examiner, Art Unit 2674
/ONEAL R MISTRY/Supervisory Patent Examiner, Art Unit 2674
Prosecution Insights