Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claims 1-37 are presented in RCE of 5/22/2026.
Amendments are entered.
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.
Claim(s) 1, 2, 8, 9, 10, 11, 17, 18, 19, 25, 26, 27, 33, 3, 12, 20 and 28 is/are rejected under 35 U.S.C. 103 as being unpatentable over in view of Chen (2019/0130580) in view of Jihye Kim et al. (US 2022/0148284) in further view of Kim – KR 2020/0095336 (prior art of record).
As to claim 1:
Chen discloses:
One or more processors comprising: one or more circuits circuitry to use one or more neural networks (¶0548, 0551, processor, 0006, 0007, neural networks) to identify one or more objects within one or more images (¶0007, 0214, 0327, identification of object(s) in one or more frames,) based, at least in part, on a selection of one or more bounding boxes from a plurality of bounding boxes, generated by the one or more neural networks and corresponding to one or more objects within one or more images, wherein the selection is made based, at least in part, on a size of the one or more objects and a ratio between an area of overlap and an overall area of the plurality of bounding boxes. (0329-0331, selecting from a set of bounding boxes comprised of foreground bounding box/high confidence bounding boxes, with an overlap ratio and the union of all reverent other boxes. ¶0022, 0272, any given high confidence bounding region (i.e. box) is selected for consideration based on when object of interest is larger than a threshold size)
Chen discloses the selection of bounding boxes also based on a bounding box removal threshold identified from a plurality of thresholds according to a size of the one or more objects as well as is silent on the mechanism for object segmentation/identification being a neural network.
However, selection of bounding boxes also based on a bounding box removal threshold identified from a plurality of thresholds is merely the well-known technique of non-maximum suppression scheme, as described in Jihye Kim , in at least ¶0049, wherein Kim, in a related field of image segmentation based on bounding boxes’ selection, discloses a segmentation process/system wherein multiple proposed bounding boxes are going through a non-maximum suppression scheme (NMS), wherein a removal threshold is used to filter said bounding boxes, i.e. removing unnecessary/repeated boxes with IOU meets said threshold. Kim also in at least ¶0063, 0047, 0130 discloses using neural network as architecture for a detector to generate bounding boxes for objects.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that Chen’s system to employ neural network to perform segmentation for object identification and that the bounding boxes selection in Chen would incorporate the well-known NMS scheme, i.e. using a removal threshold to filter unnecessary bounding boxes. This filtering implementation advantageously reduces unnecessary extra processing due to repeated/unnecessary bounding boxes are effectively removed, thus reducing waste of processing resources. Also using AI-based, specifically neural network usage allows for smart and trainable mechanism with high reliability/precision level that can be further improved through extensive trainings.
Jihye Kim discloses using a removal threshold per the discussion above, however is silent on the threshold is selectable among a plurality of thresholds corresponding to size of one or more objects.
Kim, in a related field of endeavor, in page 1, Abstract, page 5 – 6, page 15, lines 8-16, Fig. 4, discloses segmentation/classification of objects wherein objects are classified into classes, each class has a corresponding size and associated NMS threshold, thus the system has a plurality of NMS threshold to select from. NMS for filtering bounding boxes to select the correct bounding box is also used (see page 5, 15) for bounding boxes selection, and further: the system selects a NMS threshold value (i.e. removal threshold) associated with their respective target objects s among a plurality of NMS threshold values.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the module for selecting bounding boxes in Chen to be implemented using selective NMS threshold scheme based on the object’s hyperparameter. This implementation advantageously allow the system to better match different geometric behavior of different classes of objects (shape/dimensions etc..)
As to claim 10:
Chen discloses:
a computer-implemented method of determining a bounding box for an object comprising: using one or more neural networks (¶0548, 0551, processor, 0006, 0007, neural networks) to identify one or more objects within one or more images (¶0007, 0214, 0327, identification of object(s) in one or more frames,) to identify one or more objects within one or more images based, at least in part, on a selection of select one or more bounding boxes from a plurality of bounding boxes, generated by the one or more neural networks and corresponding to one or more objects within one or more images, wherein the selection is made based, at least in part, on a size of the one or more objects and a ratio between an area of overlap and an overall area corresponding to the plurality of bounding boxes. (0329-0331, selecting from a set of bounding boxes comprised of foreground bounding box/high confidence bounding boxes, using IoU method (same as Applicant’s specification) with an overlap ratio and the union of all reverent other boxes. ¶0022, 0272, any given high confidence bounding region (i.e. box) is selected for consideration based on when object of interest is larger than a threshold size)
Chen discloses the selection of bounding boxes also based on a bounding box removal threshold identified from a plurality of thresholds according to a size of the one or more objects as well as is silent on the mechanism for object segmentation/identification being a neural network.
However, selection of bounding boxes also based on a bounding box removal threshold identified from a plurality of thresholds is merely the well-known technique of non-maximum suppression scheme, as described in Jihye Kim , in at least ¶0049, wherein Kim, in a related field of image segmentation based on bounding boxes’ selection, discloses a segmentation process/system wherein multiple proposed bounding boxes are going through a non-maximum suppression scheme (NMS), wherein a removal threshold is used to filter said bounding boxes, i.e. removing unnecessary/repeated boxes with IOU meets said threshold. Kim also in at least ¶0063, 0047, 0130 discloses using neural network as architecture for a detector to generate bounding boxes for objects.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that Chen’s system to employ neural network to perform segmentation for object identification and that the bounding boxes selection in Chen would incorporate the well-known NMS scheme, i.e. using a removal threshold to filter unnecessary bounding boxes. This filtering implementation advantageously reduces unnecessary extra processing due to repeated/unnecessary bounding boxes are effectively removed, thus reducing waste of processing resources. Also using AI-based, specifically neural network usage allows for smart and trainable mechanism with high reliability/precision level that can be further improved through extensive trainings.
Jihye Kim discloses using a removal threshold per the discussion above, however is silent on the threshold is selectable among a plurality of thresholds corresponding to size of one or more objects.
Kim, in a related field of endeavor, in page 1, Abstract, page 5 – 6, page 15, lines 8-16, Fig. 4, discloses segmentation/classification of objects wherein objects are classified into classes, each class has a corresponding size and associated NMS threshold, thus the system has a plurality of NMS threshold to select from. NMS for filtering bounding boxes to select the correct bounding box is also used (see page 5, 15) for bounding boxes selection, and further: the system selects a NMS threshold value (i.e. removal threshold) associated with their respective target objects.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the module for selecting bounding boxes in Chen to be implemented using selective NMS threshold scheme based on the object’s hyperparameter. This implementation advantageously allow the system to better match different geometric behavior of different classes objects.
As to claim 18:
Chen discloses:
A non-transitory machine-readable medium having stored thereon a set of instructions, which if performed by one or more processors, cause the one or more processors to use one or more neural networks (¶0548, 0551, processor, 0006, 0007, neural networks) to identify one or more objects within one or more images (¶0007, 0214, 0327, identification of object(s) in one or more frames,) to identify one or more objects within one or more images based, at least in part, on a selection of select one or more bounding boxes from a plurality of bounding boxes, generated by the one or more neural networks and corresponding to one or more objects within one or more images, wherein the selection is made based, at least in part, on a size of the one or more objects and a ratio between an area of overlap and an overall area corresponding to the plurality of bounding boxes. (0329-0331, selecting from a set of bounding boxes comprised of foreground bounding box/high confidence bounding boxes, using IoU method (same as Applicant’s specification) with an overlap ratio and the union of all reverent other boxes. ¶0022, 0272, any given high confidence bounding region (i.e. box) is selected for consideration based on when object of interest is larger than a threshold size)
Chen discloses the selection of bounding boxes also based on a bounding box removal threshold identified from a plurality of thresholds according to a size of the one or more objects as well as is silent on the mechanism for object segmentation/identification being a neural network.
However, selection of bounding boxes also based on a bounding box removal threshold identified from a plurality of thresholds is merely the well-known technique of non-maximum suppression scheme, as described in Jihye Kim , in at least ¶0049, wherein Kim, in a related field of image segmentation based on bounding boxes’ selection, discloses a segmentation process/system wherein multiple proposed bounding boxes are going through a non-maximum suppression scheme (NMS), wherein a removal threshold is used to filter said bounding boxes, i.e. removing unnecessary/repeated boxes with IOU meets said threshold. Kim also in at least ¶0063, 0047, 0130 discloses using neural network as architecture for a detector to generate bounding boxes for objects.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that Chen’s system to employ neural network to perform segmentation for object identification and that the bounding boxes selection in Chen would incorporate the well-known NMS scheme, i.e. using a removal threshold to filter unnecessary bounding boxes. This filtering implementation advantageously reduces unnecessary extra processing due to repeated/unnecessary bounding boxes are effectively removed, thus reducing waste of processing resources. Also using AI-based, specifically neural network usage allows for smart and trainable mechanism with high reliability/precision level that can be further improved through extensive trainings.
Jihye Kim discloses using a removal threshold per the discussion above, however is silent on the threshold is selectable among a plurality of thresholds corresponding to size of one or more objects.
Kim, in a related field of endeavor, in page 1, Abstract, page 5 – 6, page 15, lines 8-16, Fig. 4, discloses segmentation/classification of objects wherein objects are classified into classes, each class has a corresponding size and associated NMS threshold, thus the system has a plurality of NMS threshold to select from. NMS for filtering bounding boxes to select the correct bounding box is also used (see page 5, 15) for bounding boxes selection, and further: the system selects a NMS threshold value (i.e. removal threshold) associated with their respective target object.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the module for selecting bounding boxes in Chen to be implemented using selective NMS threshold scheme based on the object’s hyperparameter. This implementation advantageously allow the system to better match different geometric behavior of different classes of objects.
As to claim 26:
Chen discloses:
A system comprising: one or more processors; and computer-readable media having stored thereon executable instructions that, as a result of being performed by the one or more processors, cause the system to use one or more neural networks (¶0548, 0551, processor, 0006, 0007, neural networks) to identify one or more objects within one or more images (¶0007, 0214, 0327, identification of object(s) in one or more frames,) to identify one or more objects within one or more images based, at least in part, on a selection of select one or more bounding boxes from a plurality of bounding boxes, generated by the one or more neural networks and corresponding to one or more objects within one or more images, wherein the selection is made based, at least in part, on a size of the one or more objects and a ratio between an area of overlap and an overall area corresponding to the plurality of bounding boxes. (0329-0331, selecting from a set of bounding boxes comprised of foreground bounding box/high confidence bounding boxes, using IoU method (same as Applicant’s specification) with an overlap ratio and the union of all reverent other boxes. ¶0022, 0272, any given high confidence bounding region (i.e. box) is selected for consideration based on when object of interest is larger than a threshold size)
Chen discloses the selection of bounding boxes also based on a bounding box removal threshold identified from a plurality of thresholds according to a size of the one or more objects as well as is silent on the mechanism for object segmentation/identification being a neural network.
However, selection of bounding boxes also based on a bounding box removal threshold identified from a plurality of thresholds is merely the well-known technique of non-maximum suppression scheme, as described in Jihye Kim , in at least ¶0049, wherein Kim, in a related field of image segmentation based on bounding boxes’ selection, discloses a segmentation process/system wherein multiple proposed bounding boxes are going through a non-maximum suppression scheme (NMS), wherein a removal threshold is used to filter said bounding boxes, i.e. removing unnecessary/repeated boxes with IOU meets said threshold. Kim also in at least ¶0063, 0047, 0130 discloses using neural network as architecture for a detector to generate bounding boxes for objects.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that Chen’s system to employ neural network to perform segmentation for object identification and that the bounding boxes selection in Chen would incorporate the well-known NMS scheme, i.e. using a removal threshold to filter unnecessary bounding boxes. This filtering implementation advantageously reduces unnecessary extra processing due to repeated/unnecessary bounding boxes are effectively removed, thus reducing waste of processing resources. Also using AI-based, specifically neural network usage allows for smart and trainable mechanism with high reliability/precision level that can be further improved through extensive trainings.
Jihye Kim discloses using a removal threshold per the discussion above, however is silent on the threshold is selectable among a plurality of thresholds corresponding to size of one or more objects.
Kim, in a related field of endeavor, in page 1, Abstract, page 5 – 6, page 15, lines 8-16, Fig. 4, discloses segmentation/classification of objects wherein objects are classified into classes, each class has a corresponding size and associated NMS threshold, thus the system has a plurality of NMS threshold to select from. NMS for filtering bounding boxes to select the correct bounding box is also used (see page 5, 15) for bounding boxes selection, and further: the system selects a NMS threshold value (i.e. removal threshold) associated with their respective target objects.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the module for selecting bounding boxes in Chen to be implemented using selective NMS threshold scheme based on the object’s hyperparameter. This implementation advantageously allow the system to better match different geometric behavior of different classes of objects.
As to claims 2, 11, 19 and 27:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations of claim 1/10/18/26, wherein: an individual bounding box is not selected as a result of a ratio of the individual bounding box the one or more bounding boxes being greater than a removal threshold value. (Jihye Kim , ¶0049, “removing overlapping proposals if the overlapping proposals are greater than or equal to a threshold may be repeated, and a plurality of overlapping output results for the same target may be deleted”)
As to claims 8, 17, 25, 33:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations of claim 1/10/18/26, wherein the one or more images include a plurality of objects and the selected one or more bounding boxes includes a bounding box for each of the plurality of objects. (See Chen, ¶0013, Kim, page 6, lines 26-27, a plurality of boxes, selecting each box corresponding to each object)
As to claim 9:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations of claim 1, wherein: the one or more neural networks detect the one or more objects; and the one or more neural networks generate the plurality of bounding boxes. (See Chen, ¶0013, Kim, page 5, lines 38 through line 5 of page 6, detecting the one or more objects. Page 6, lines 9-13, 27, a plurality of boxes generated.)
As to claims 3, 12, 20 and 28:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations of claim 1/10/18/26, Chen disclose the size of the one or more objects is determined based on a bounding box associated with the one or more objects. (See Chen, ¶0252, 0272)
Claim(s) 4, 13, 21 and 29 is/are rejected under 35 U.S.C. 103 as being unpatentable over in view of Chen (2019/0130580) in view of Jihye Kim et al. (US 2022/0148284) in further view of Kim – KR 2020/0095336 (prior art of record)and in view of Wang et al. (WO 2018140062).
As to claims 4, 13, 21 and 29:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations of claim 1/10/18/26, however is silent on the size of the one or more objects is determined based at least in part on a distance between the one or more objects and a camera used to obtain the one or more images.
Such description serves as narrative description in passive tense on how a size of an object can be determined, and not necessarily tied to the claimed process/system.
Regardless, to address this limitation, Wang, in a related field of endeavor, also discloses in Abstract, page 8, lines 15-35 a system/method for object detection, in which a size of an object is determined based on the distance of the object from camera (See at least ¶0009, ¶0026, determining size of object based on object-camera distance)
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that size of object in Kim can be reasonably estimated via camera distance in manner disclosed by Wang. As Wang demonstrates a reliable mathematical relationship between such distance and dimension of an object in ¶0026, such computation at least serves as a reasonable estimate to determine at the size of the object.
Claim(s) 7, 16, 24 and 32 is/are rejected under 35 U.S.C. 103 as being unpatentable over in view of Chen (2019/0130580) in view of Jihye Kim et al. (US 2022/0148284) in further view of Kim – KR 2020/0095336 (prior art of record) and in view of Boykin et al. (US 2018/0025636).
As to claims 7, 24:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations of claim 1/18, however is silent on: the one or more objects includes a first vehicle; and the one or more images are obtained from a camera/imaging device on a second vehicle.
Boykin discloses a system and method for object detection and monitoring in an image analysis model, specifically in Fig. 15 and ¶0085 and 0086, which is implemented with a camera in a monitoring vehicle and the object being another vehicle on the street.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the system of object detection of Kim can be incorporated in vehicle detection system in an in-vehicle system as described in Boykin. Limitation of claim 7 merely describes a specific application of the system of claim 1. Boykin has shown that such object detection system can be incorporated in vehicle monitoring system in ¶0086, which is helpful in indicating a distracted or intoxicated driver of another vehicle.
As to claim 16:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations of claim 10, however is silent on: the one or more objects includes a person; and the one or more images are obtained from a camera on a vehicle.
Boykin discloses a system and method for object detection and monitoring in an image analysis model, specifically in Fig. 16 and associated texts, which is implemented with a camera in a monitoring vehicle and the object being another vehicle or a person on the street.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the system of object detection of Kim can be incorporated in vehicle detection system in an in-vehicle system as described in Boykin. Limitation of claim 7 merely describes a specific application of the system of claim 1. Boykin has shown that such object detection system can be incorporated in vehicle monitoring system in ¶0086, which is helpful in avoiding collision with pedestrians.
As to claim 32:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations of claim 26, however is silent on: the one or more images are obtained from a camera mounted on an autonomous vehicle.
Boykin discloses a system and method for object detection and monitoring in an image analysis model, specifically in ¶0074, the one or more images are obtained from a camera mounted on an autonomous vehicle.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the system of object detection of Kim can be incorporated in vehicle/person detection system in an in-vehicle system of an autonomous vehicle as described in Boykin. Limitation of claim 7 merely describes a specific application of the system of claim 1. Boykin has shown that such object detection system can be incorporated in vehicle monitoring system in ¶0074 which can be helpful in monitoring or pursuing vehicle or pedestrians of interest.
Claim(s) 34-37 is/are rejected under 35 U.S.C. 103 as being unpatentable over in view of Chen (2019/0130580) in view of Jihye Kim et al. (US 2022/0148284) in further view of Kim – KR 2020/0095336 (prior art of record) in view of He (CN 111626208A).
As to claim 34, 35, 36 and 37:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations of claims 1, 10, 18 and 26 respectively, wherein a bounding box for the individual is selected from bounding boxes that have not been removed (Jihye Kim , ¶0049, and in particular, “removing overlapping proposals if the overlapping proposals are greater than or equal to a threshold may be repeated, and a plurality of overlapping output results for the same target may be deleted”, naturally, one cannot choose a box that is already deleted)
Regarding: the bounding box removal threshold for an individual object of the one or more objects is selected from the plurality of thresholds based, at least in part, on a size of the individual object. Kim discloses selections of a removal threshold from a plurality of thresholds associated with objects having minimum size. While the choice of threshold based on size is not expressively stated, however it is well known NMS threshold is dimension-based and thus is proportional to object size.
Indeed, He discloses a system and method for object detection wherein the system selects a particular suitable IoU threshold that is according to the target size (Page 7, lines 1-15).
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the system of object detection of Chen’s combination can be incorporated using adjustable IOU threshold. Given that Kim also employs a bound box selection threshold (page 6, line 26-28), the implementation of Ma is well-fit. This implementation advantageously allow the system to better match different geometric behavior of large vs. small objects. With larger object classes having different threshold than smaller classes, this results to fewer missed detection, fewer residual boxes on large objects and overall better call precision on box selection.’
Claim(s) 5, 6, 14, 15, 22, 23, 30, 31 is/are rejected under 35 U.S.C. 103 as being unpatentable over in view of Chen (2019/0130580) in view of Jihye Kim et al. (US 2022/0148284) in further view of Kim – KR 2020/0095336 (prior art of record) and in view of Elachqar et al. (US 2022/0188541).
As to claims 6, 23, 31:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations claim 1/18/26, and discloses the selection by performing non-maximum suppression on the plurality of bounding boxes (See Kim, Page 6, lines 22-25, performing non-maximum suppression for selection of bounding box), however does not delves into the detail of such filtering process, namely: with respect to a confidence measure associated with each bounding box of the plurality of bounding boxes.
Elachgar, in a related field of endeavor, also discloses a system/method for bounding box selection in which non-maximum suppression is used. Elachgar provides a more details view into the non-maximum algorithm in at least ¶0079, in which the algorithm compare each confidence level of each of the plurality of bounding boxes with the highest confidence level box, and remove any boxes that are deemed negligible.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the non-maximum suppression algorithm in Kim comprises comparison each confidence level of each of the plurality of bounding boxes with the highest confidence level box, and remove any boxes that are deemed negligible, as such steps are the definition of the algorithm is about. The implementation naturally eliminates negligible boxes, reducing the size of the set of bounding boxes, thus reducing the needs (computation resource saving) for the entire process.
As to claims 5, 14, 22, 30:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations claim 1/10/18/26, while Kim does not explicitly disclose each bounding box in the plurality of bounding boxes has an associated confidence measure; and a bounding box for an individual object is selected based at least in part on a confidence measure associated with the bounding box.
However, such description is part of an algorithm called non-maximum suppression algorithm, as disclosed in Kim (See Page 6, lines 22-25, performing non-maximum suppression for selection of bounding box)
To clarify this assertion, Elachgar, in a related field of endeavor, also discloses a system/method that provides a more details view into the non-maximum algorithm in at least ¶0079, in which the algorithm compare each respective confidence level of each of the plurality of bounding boxes with the highest confidence level box, and remove any boxes that are deemed negligible as part of the selection process. The selection process is thus based on the set of remaining boxes of highest confidence level resulted from this algorithm.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the non-maximum suppression algorithm in Kim comprises comparison each confidence level of each of the plurality of bounding boxes with the highest confidence level box, and remove any boxes that are deemed negligible, as such steps are the definition of the algorithm is about. The implementation naturally eliminates negligible boxes, reducing the size of the set of bounding boxes, thus reducing the needs (computation resource saving) for the entire process.
As to claim 15:
Chen in view of Jihye Kim and in further view of Kim discloses all limitations claim 10, and discloses the one or more bounding boxes is selected by performing non-maximum suppression on the plurality of bounding boxes (See Page 6, lines 22-25, performing non-maximum suppression for selection of bounding box), however does not delves into the detail of such filtering process, namely: the one or more bounding boxes is selected by filtering the plurality of bounding boxes with respect to a confidence measure associated with each bounding box of the plurality of bounding boxes.
Elachgar, in a related field of endeavor, also discloses a system/method for bounding box selection in which non-maximum suppression is used. Elachgar provides a more details view into the non-maximum algorithm in at least ¶0079, in which the algorithm compare each confidence level of each of the plurality of bounding boxes with the highest confidence level box, and remove any boxes that are deemed negligible, thus “filtering”.
It would have been obvious to one of ordinary skill in the art before the effective filing time of the invention that the non-maximum suppression algorithm in Kim comprises comparison each confidence level of each of the plurality of bounding boxes with the highest confidence level box, and remove any boxes that are deemed negligible, as such steps are the definition of the algorithm is about. The implementation naturally eliminates negligible boxes, reducing the size of the set of bounding boxes, thus reducing the needs (computation resource saving) for the entire process.
Response to Arguments
Applicant’s arguments with respect to claim(s) have been considered but are moot because the new ground of rejection.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
US 2022/0413464 - Examples of methods for registering objects are described herein. In some examples, a method includes determining a set of overlap scores based on a set of orientations between a first bounding box of a three-dimensional (3D) object model and a second bounding box of a 3D scan of an object. In some examples, the method includes registering the 3D scan with the 3D object model based on the set of overlap scores.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to QUAN M HUA whose telephone number is (571)270-7232. The examiner can normally be reached 10:30-6:30.
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, Anthony Addy can be reached at 571-272-7795. 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.
/QUAN M HUA/Primary Examiner, Art Unit 2645
Prosecution Insights