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 .
Response to Amendment
Applicant's response to the last Office Action, filed on 12/29/2025 has been entered and made of record.
Response to Arguments
Applicant's arguments with respect to claims 1, 10, and 11 have been considered but are moot in view of the new grounds of rejection.
Information Disclosure Statement
The information disclosure statement (IDS) filed on 10/1/2025 was considered and placed on the file of record by the examiner.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 3, 4, 5, 6, 10, 11 are rejected under 35 U.S.C. 103 as being unpatentable over Wu et al. (US 2022/0012890) in view of Md et al., “SiamGauss: Siamese region proposal network with Gaussian head for visual object tracking.”
Regarding claim 1, Wu teaches a method of object tracking based on deep-learning performed in an electronic apparatus, the method comprising:
pre-training a model for object tracking based on pre-input learning data (see para. 0223, Wu discusses pre-training and the fine tuning of the whole network, the network is trained with Kullback-Leibler divergence (KLD) loss);
receiving a target image of which at least one area contains an image corresponding an object for tracking and a search image of which at least one area contains an image corresponding the object for tracking (see para. 0223, Wu discusses pre-training and the fine tuning of the whole network, the network is trained with Kullback-Leibler divergence (KLD) loss); and
wherein the area corresponding to the object for tracking is defined by a Gaussian distribution model, wherein the information on area for tracking includes parameter values of a plurality of parameters based Gaussian distribution corresponding to the area corresponding to the object for tracking (see figure 16, para. 0078, Wu discusses converts the discrete probability map P to a Gaussian parameterization. The Gaussian parameterization G is then fed into the trained smoothing block, which incorporate the learned surface smoothness priors to generate the optimal target surface).
Wu does not expressly disclose obtaining information on area for tracking regarding to the area corresponding to the object for tracking in the search image by applying the model for object tracking, wherein the model for object tracking includes a Siamese Region Proposal Network (RPN) structure and a refining module configured to perform a rotated region of interest (ROI) aligning operation to generate a Gaussian prediction.
However, Md teaches obtaining information on area for tracking regarding to the area corresponding to the object for tracking in the search image by applying the model for object tracking (see section 2.2, Md discusses object tracking in search images by applying a model),
wherein the model for object tracking includes a Siamese Region Proposal Network (RPN) structure and a refining module configured to perform a rotated region of interest (ROI) aligning operation to generate a Gaussian prediction (see figure 2, figure 3, section 3, Md discusses object tracking includes a Siamese Region Proposal Network and refining module with asymmetric filters that generate Gaussian predictions).
Motivation to combine may be gleaned from the prior art considered. It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to modify the invention of Wu with Md to derive at the invention of claim 1. The result would have been expected, routine, and predictable in order to perform Gaussian object detection and tracking.
The determination of obviousness is predicated upon the following: One skilled in the art would have been motivated to modify Wu in this manner in order to improve Gaussian object detection and tracking by applying a Siamese region proposal network model that focuses on a search region. Furthermore, the prior art collectively includes each element claimed (though not all in the same reference), and one of ordinary skill in the art could have combined the elements in this manner explained using known engineering design, interface and/or programming techniques, without changing a fundamental operating principle of Wu, while the teaching of Md continues to perform the same function as originally taught prior to being combined, in order to produce the repeatable and predictable result of using a model that implements a target search region to properly detect and track objects. The Wu and Md systems perform object detection in images, therefore one of ordinary skill in the art would have reasonable expectation of success in the combination. It is for at least the aforementioned reasons that the examiner has reached a conclusion of obviousness with respect to the claim in question.
Regarding claim 3, Wu teaches wherein the pre-training the model for object tracking further includes updating parameters corresponding to at least one layer included in the model for object tracking in a direction of minimizing a result value of a loss function (see para. 0223, Wu discusses performing backpropagation update parameters to minimize the Kullback-Leibler divergence (KLD) loss),
wherein the loss function is set to indicate a difference between the information on area for tracking obtained by applying the model for object tracking and a ground truth included in the learning data (see para. 0223, Wu discusses pre-training and the fine tuning of the whole network, the network is trained with Kullback-Leibler divergence (KLD) loss).
The same motivation of claim 1 is applied to claim 3. Motivation to combine may be gleaned from the prior art considered. It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to modify the invention of Wu with Md to derive at the invention of claim 3. The result would have been expected, routine, and predictable in order to perform Gaussian object detection and tracking.
Regarding claim 4, Wu teaches wherein the loss function is defined based on a Kullback-Leibler divergence value between Gaussian distribution corresponding to the information on area for tracking and Gaussian distribution corresponding to the ground truth (see para. 0223, Wu discusses pre-training and the fine tuning of the whole network, the network is trained with Kullback-Leibler divergence (KLD) loss).
The same motivation of claim 1 is applied to claim 4. Motivation to combine may be gleaned from the prior art considered. It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to modify the invention of Wu with Md to derive at the invention of claim 4. The result would have been expected, routine, and predictable in order to perform Gaussian object detection and tracking.
Regarding claim 5, Md teaches wherein the RPN includes an area proposal module for obtaining candidate area information corresponding to the at least one area of the search image based on at least one convolution layer for obtaining a feature map from the target image and the search image and the feature maps (see section 3, Md discusses a Siamese region proposal network, wherein captured images are localized to the search image using one or more target positions); and
the model for object tracking includes a refining module including at least one fully connected layer for obtaining the information on area for tracking from the feature maps and the candidate area information obtained from the RPN (see section 3, Md discusses a Siamese region proposal network, wherein captured images are localized to the search image using one or more target positions).
The same motivation of claim 1 is applied to claim 5. Motivation to combine may be gleaned from the prior art considered. It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to modify the invention of Wu with Md to derive at the invention of claim 5. The result would have been expected, routine, and predictable in order to perform Gaussian object detection and tracking.
Regarding claim 6, Md teaches wherein the Siamese RPN structure includes a first branch that receives the target image as an input and includes at least one convolution layer and a second branch that receives the search image as an input and includes at least one convolution layer that share parameters with at least one layer of the first branch (see figure 2, figure 3, section 3, Md discusses RPN includes a Siamese-network structure including template image and search image, and implementing convolutional layers),
wherein, based on a cross-correlation operation performed based on a first feature map obtained from the at least one convolution layer of the first branch and a second feature map obtained from the at least one convolution layer of the second branch, the area proposal module obtains a Gaussian feature map including information on a Gaussian area and a class feature map that contains information about a score corresponding to at least one anchor that is set in each of at least one area of the search image (see figure 2, figure 3, section 3, Md discusses cross-correlation operation, Gaussian map and feature maps with classification scores),
the method further comprising obtaining candidate area information based on a score for each anchor identified based on the class feature map and the Gaussian feature map (see figure 2, figure 3, section 3, Md discusses class feature map and Gaussian distribution map).
The same motivation of claim 1 is applied to claim 6. Motivation to combine may be gleaned from the prior art considered. It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to modify the invention of Wu with Md to derive at the invention of claim 6. The result would have been expected, routine, and predictable in order to perform Gaussian object detection and tracking.
Claim 10 is rejected as applied to claim 1 as pertaining to a corresponding apparatus.
Claim 11 is rejected as applied to claim 1 as pertaining to a corresponding computer-readable non-transitory recording medium.
Claims 2, 9 are rejected under 35 U.S.C. 103 as being unpatentable over Wu et al. (US 2022/0012890) in view of Md et al., “SiamGauss: Siamese region proposal network with Gaussian head for visual object tracking” in view of Llerena et al., “Gaussian Bounding Boxes and Probabilistic Intersection-over-Union for Object Detection”.
Regarding claim 2, Wu and Md do not expressly disclose wherein the plurality of parameters include at least parameters related to position coordinates of the Gaussian distribution, parameters related to a shape of the Gaussian distribution and parameters related to a rotation angle of the Gaussian distribution. However, Llerena teaches wherein the plurality of parameters include at least parameters related to position coordinates of the Gaussian distribution, parameters related to a shape of the Gaussian distribution and parameters related to a rotation angle of the Gaussian distribution (see abstract, section 3, Llerena discusses calculating Gaussian distributions of objects in images and generating Gaussian Bounding Boxes that represent object shape and location, the Gaussian distribution contain parameters related to the shape and rotation angle).
Motivation to combine may be gleaned from the prior art considered. It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to modify the invention of Wu and Md with Llerena to derive at the invention of claim 2. The result would have been expected, routine, and predictable in order to perform Gaussian object detection and tracking.
The determination of obviousness is predicated upon the following: One skilled in the art would have been motivated to modify Wu and Md in this manner in order to improve Gaussian object detection and tracking by calculating the Gaussian distribution of an object and generate Gaussian bounding boxes based on shape and rotation angle parameters, that improves object shape segmentation and similarity comparison for object tracking. Furthermore, the prior art collectively includes each element claimed (though not all in the same reference), and one of ordinary skill in the art could have combined the elements in this manner explained using known engineering design, interface and/or programming techniques, without changing a fundamental operating principle of Wu and Md, while the teaching of Llerena continues to perform the same function as originally taught prior to being combined, in order to produce the repeatable and predictable result of calculating a Gaussian distribution with shape and rotation angle parameters to improve object segmentation and similarity calculations between images. The Wu, Md, and Llerena systems perform object detection in images, therefore one of ordinary skill in the art would have reasonable expectation of success in the combination. It is for at least the aforementioned reasons that the examiner has reached a conclusion of obviousness with respect to the claim in question.
Regarding claim 9, Wu and Md do not expressly disclose wherein the search image includes a first frame image included in first video containing a plurality of frames and a second frame image after the first frame image, the method further comprising: obtaining first information on area for tracking based on the first frame image; obtaining second information on area for tracking based on the second frame image; modifying parameter values related to a shape and parameter values related to a rotation angle of an area corresponding to the object for tracking in the second information on area for tracking based on the first information on area for tracking.
However, Llerena teaches wherein the search image includes a first frame image included in first video containing a plurality of frames and a second frame image after the first frame image, the method further comprising: obtaining first information on area for tracking based on the first frame image (see section 3, section 4.3, Llerena discusses object detection over multiple images); obtaining second information on area for tracking based on the second frame image (see section 3, section 4.3, Llerena discusses object detection over multiple images); modifying parameter values related to a shape and parameter values related to a rotation angle of an area corresponding to the object for tracking in the second information on area for tracking based on the first information on area for tracking (see abstract, section 3, Llerena discusses calculating Gaussian distributions of objects in images and generating Gaussian Bounding Boxes that represent object shape and location, the Gaussian distribution contain parameters related to the shape and rotation angle. The gaussian bounding boxes are used to detect and compare objects in different images).
Motivation to combine may be gleaned from the prior art considered. It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to modify the invention of Wu and Md with Llerena to derive at the invention of claim 9. The result would have been expected, routine, and predictable in order to perform Gaussian object detection and tracking.
The determination of obviousness is predicated upon the following: One skilled in the art would have been motivated to modify Wu and Md in this manner in order to improve Gaussian object detection and tracking by calculating the Gaussian distribution of an object and generate Gaussian bounding boxes based on shape and rotation angle parameters, that improves object shape segmentation and similarity comparison for object tracking. Furthermore, the prior art collectively includes each element claimed (though not all in the same reference), and one of ordinary skill in the art could have combined the elements in this manner explained using known engineering design, interface and/or programming techniques, without changing a fundamental operating principle of Wu and Md, while the teaching of Llerena continues to perform the same function as originally taught prior to being combined, in order to produce the repeatable and predictable result of calculating a Gaussian distribution with shape and rotation angle parameters to improve object segmentation and similarity calculations between images. The Wu, Md, and Llerena systems perform object detection in images, therefore one of ordinary skill in the art would have reasonable expectation of success in the combination. It is for at least the aforementioned reasons that the examiner has reached a conclusion of obviousness with respect to the claim in question.
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Wu et al. (US 2022/0012890) in view of Md et al., “SiamGauss: Siamese region proposal network with Gaussian head for visual object tracking” in view of Lee et al. (US 2023/0316536).
Regarding claim 7, Wu and Md do not expressly disclose wherein the refining module receives the candidate area information and the feature map obtained from the RPN as inputs, and obtains a feature map corresponding to a candidate area based on a rotated region of interest (ROI) aligning operation based on bilinear interpolation, and the refining module is configured to obtain the information on area for tracking based on the feature map corresponding to the candidate area.
However, Lee teaches wherein the refining module receives the candidate area information and the feature map obtained from the RPN as inputs, and obtains a feature map corresponding to a candidate area based on a rotated region of interest (ROI) aligning operation based on bilinear interpolation (see figure 3, para. 0064-0065, Lee discusses receiving feature data from RPNs and performing bilinear interpolation to align feature data), and
the refining module is configured to obtain the information on area for tracking based on the feature map corresponding to the candidate area (see figure 3, para. 0065, 0070, Lee discusses object tracking based on bounding boxes that identify regions of classified objects in the image).
Motivation to combine may be gleaned from the prior art considered. It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to modify the invention of Wu and Md with Lee to derive at the invention of claim 7. The result would have been expected, routine, and predictable in order to perform Gaussian object detection and tracking.
The determination of obviousness is predicated upon the following: One skilled in the art would have been motivated to modify Wu and Md in this manner in order to improve Gaussian object detection and tracking by performing bilinear interpolation to align feature maps in a region proposal network model. Furthermore, the prior art collectively includes each element claimed (though not all in the same reference), and one of ordinary skill in the art could have combined the elements in this manner explained using known engineering design, interface and/or programming techniques, without changing a fundamental operating principle of Wu and Md, while the teaching of Lee continues to perform the same function as originally taught prior to being combined, in order to produce the repeatable and predictable result of performing bilinear interpolation to align feature maps to properly detect and track objects. The Wu, Md, and Lee systems perform object detection in images, therefore one of ordinary skill in the art would have reasonable expectation of success in the combination. It is for at least the aforementioned reasons that the examiner has reached a conclusion of obviousness with respect to the claim in question.
Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Wu et al. (US 2022/0012890) in view of Md et al., “SiamGauss: Siamese region proposal network with Gaussian head for visual object tracking” in view of Li et al. (US 2021/0295529).
Regarding claim 8, Wu and Md do not expressly disclose wherein the information on area for tracking corresponding to the object for tracking includes a plurality of parameter values for a plurality of areas and information about confidence corresponding to each of the plurality of areas,
the method further comprising obtaining a weighted-mean value of the plurality of parameter vales based on the information about the confidence.
However, Li teaches wherein the information on area for tracking corresponding to the object for tracking includes a plurality of parameter values for a plurality of areas and information about confidence corresponding to each of the plurality of areas, the method further comprising obtaining a weighted-mean value of the plurality of parameter vales based on the information about the confidence (see para. 0099, Li discusses weighted mean of the possibility parameters associated with the object boxes).
Motivation to combine may be gleaned from the prior art considered. It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to modify the invention of Wu and Md with Li to derive at the invention of claim 8. The result would have been expected, routine, and predictable in order to perform Gaussian object detection and tracking.
The determination of obviousness is predicated upon the following: One skilled in the art would have been motivated to modify Wu and Md in this manner in order to improve Gaussian object detection and tracking by calculating weighted mean values associated with object boxes. Furthermore, the prior art collectively includes each element claimed (though not all in the same reference), and one of ordinary skill in the art could have combined the elements in this manner explained using known engineering design, interface and/or programming techniques, without changing a fundamental operating principle of Wu and Md, while the teaching of Li continues to perform the same function as originally taught prior to being combined, in order to produce the repeatable and predictable result of determining weighted mean values to properly detect and track objects. The Wu, Md, and Li systems perform object detection in images, therefore one of ordinary skill in the art would have reasonable expectation of success in the combination. It is for at least the aforementioned reasons that the examiner has reached a conclusion of obviousness with respect to the claim in question.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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.
Contact Information
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KENNY A CESE whose telephone number is (571) 270-1896. The examiner can normally be reached on Monday – Friday, 9am – 4pm.
If attempts to reach the primary examiner by telephone are unsuccessful, the examiner’s supervisor, Gregory Morse can be reached on (571) 272-3838. 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 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/Kenny A Cese/
Primary Examiner, Art Unit 2663