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 .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on July 7, 2025 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement has been considered by the examiner.
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, 10, 11, 13 and 20 are rejected under 35 U.S.C 103 as being unpatentable over Pan et al. US Patent Publication No. US-20190172184-A1 (hereinafter Pan) in view of Nicolas ‘Turbulence mitigation in imagery including moving objects from a static event camera’ (hereinafter Nicolas), Dias ‘Semantic Segmentation refinement’ (hereinafter Dias) and Longlong ‘Coarse-to-Fine Semantic Segmentation From Image-Level Labels’ (hereinafter Longlong) and further in view of Xuelian ‘Implicit Motion Handling for Video Camouflaged Object Detection’ (hereinafter Xuelian).
Regarding claim 1, Pan discloses a method of using an unsupervised region-growing network to perform object segmentation for video data degraded by atmospheric turbulence (Pan in [0001] discloses, “The present invention relates to methods of video stabilization and artefacts removal for turbulence effect compensation in long distance imaging”), comprising: obtaining input data including the video data degraded by the atmospheric turbulence (Pan in [0014] discloses, “there is provided a method of correcting for a turbulence effect in a video comprising a plurality of frames”); extracting a video frame sequence from the video data (Pan in [0063] discloses, “The method 200 then proceeds from sub-process 240 to decision step 250 . Step 250 determines whether there are more raw turbulence video frames of the video to be processed. If there are more turbulence video frames to be processed (YES), the method 200 proceeds from step 250 to step 210. Therefore, sub-processes 210 to 240 are performed for each turbulence video frame”).
Pan doesn’t disclose the following limitation as further recited in the claim.
Nicolas discloses training, by a computer, an (Nicolas in [Section – 3.4, Paragraph – 1] discloses, “The mask derived from event statistics and the mask derived from the error between the background and the current frame are combined using a region growing algorithm. Starting from a seed (the event-derived mask), the algorithm iteratively integrates neighboring blocks if they are marked in the error mask. This strategy makes it possible to minimize the amount of false positive (MSE outliers due to strong turbulence incorrectly classified as the moving object) and false negative (flat zones of the moving object incorrectly classified as background)”. Furthermore, Nicolas in [Page – 17, Paragraph – 1] discloses, “it helps in the segmentation of the moving object without being limited by the comparison of two integrated frames such as in an optical flow-based algorithm. Second, it improves the reconstruction of the moving object appearance and the refinement of its boundaries” wherein refinement equates to grouping loss function): generating pixel-level masks for any moving object identified within the video frame sequence (Nicolas in [Section – 3.4, Paragraph – 1] discloses, “As shown in Fig. 6, the distinction between background and a moving object is implemented using a binary mask that indicates the pixels corresponding to the moving object); applying the region-growing algorithm to generate a coarse mask for each moving object from the pixel-level masks generated for any moving object identified within the video frame sequence (Nicolas in [Section – 3.4, Paragraph – 1] discloses, “In our algorithm, a coarse mask is computed by splitting the intensity frame into subblocks (8 × 8 pixels) ... The mask derived from event statistics and the mask derived from the error between the background and the current frame are combined using a region growing algorithm. Starting from a seed (the event-derived mask), the algorithm iteratively integrates neighboring blocks if they are marked in the error mask); and outputting, using the trained RGN, the refined masks as object segmentation data for the video data received (Nicolas in [Section – 3.4, Paragraph – 1] discloses, “As shown in Fig. 6, the distinction between background and a moving object is implemented using a binary mask that indicates the pixels corresponding to the moving object. ... The mask derived from event statistics and the mask derived from the error between the background and the current frame are combined using a region growing algorithm).
It would have been obvious to one of ordinary skill in art before the effective filling date of the claimed invention to integrate the technique of Nicolas into the system of Pan because the region growing algorithm would allow the system to start from reliable motion information and expand outward to cover the moving object.
Pan and Nicolas in the combination don’t disclose the following limitation as further recited in the claim.
Dias discloses unsupervised Region-Growing Network (RGN) (Dias in [Page – 3, Paragraph – 3] discloses, “we propose the Region Growing Refinement (RGR) algorithm, an unsupervised and easily generalizable post-processing module that performs appearance-based region growing to refine the predictions generated by a CNN for semantic segmentation”).
It would have been obvious to one of ordinary skill in art before the effective filling date of the claimed invention to integrate the technique of Dias into the system of Pan in view of Nicolas because the system would not need any manually labeled data and reduce the cost and increase efficiency.
Longlong discloses applying the grouping loss function to the coarse mask generated for each moving object to generate refined masks consistent across consecutive frames of the video frame sequence corresponding to each coarse mask (Longlong in [Section – 3, Paragraph – 3] discloses, “the framework in [20] is a CNN-based network that trained with millions of unlabeled images and achieves the state-of-the-art in unsupervised object segmentation. Moreover, the student network, an 8-layer CNN trained on large scale video frames”. Furthermore, Longlong in [Section – 3(C), Paragraph – 2] also discloses, “A graph cut-based optimization method is run to minimizing the energy function that prefers connected regions having the same label. By repeating the two-step procedure until it converges, the enhanced coarse masks are obtained”. Longlong in [Section – 3(B), Paragraph – 1] discloses, “Many methods can generate the coarse masks ... Unsupervised Foreground Segmentation (UFS) [20], and Unsupervised Object Segmentation ... These methods can segment moving objects in videos or generate saliency maps for images”. The system being able to Segment and process a video equates to consecutive frames).
It would have been obvious to one of ordinary skill in art before the effective filling date of the claimed invention to integrate the technique of Longlong into the system of Pan in view of Nicolas and Dias because it would allow the system to produce more precise segments.
Pan, Nicolas, Dias and Longlong in the combination don’t disclose the following limitation as further recited in the claim.
Xuelian discloses for each of a plurality of reference frames within the video frame sequence, computing a bidirectional optical flow sequence between the reference frame and any neighboring frame of the reference frame within the video frame sequence (Xuelian in [Page – 11, Paragraph – 2] discloses, “Given five consecutive frames {It, It+1,It+2,It+3,It+4} and labelled ground-truth gtt, we first estimate forward and backward optical flow fields between frame It and It+n,n ∈ [1,4]”. Xuelian in [Page – 11, Paragraph – 5] discloses, “Bidirectional Consistency Check. To identify valid masks, we adopt forward-backward consistency check to eliminate inconsistent regions”).
It would have been obvious to one of ordinary skill in art before the effective filling date of the claimed invention to integrate the technique of Xuelian into the system of Pan view of Nicolas, Dias and Longlong because it would allow the system to be more accurate in temporal consistency.
Summary of Citations (Dias)
[Page – 3, Paragraph – 3]; “we propose the Region Growing Refinement (RGR) algorithm, an unsupervised and easily generalizable post-processing module that performs appearance-based region growing to refine the predictions generated by a CNN for semantic segmentation”.
Summary of Citations (Longlong)
[Section – 3, Paragraph – 1]; “a trained 8-layer CNN is employed to generate the initial coarse masks for images. Secondly, a graph-based model is employed to enhance the quality of the initial coarse masks based on the object prior. Finally, these enhanced masks together with the input images and their category labels are used to recursively train a fully convolution network designed for semantic segmentation”.
[Section – 3(B), Paragraph – 1]; “Many methods can generate the coarse masks ... Unsupervised Foreground Segmentation (UFS) [20], and Unsupervised Object Segmentation ... These methods can segment moving objects in videos or generate saliency maps for images”.
[Section – 3, Paragraph – 3]; “the framework in [20] is a CNN-based network that trained with millions of unlabeled images and achieves the state-of-the-art in unsupervised object segmentation. Moreover, the student network, an 8-layer CNN trained on large scale video frames”.
[Section – 3(C), Paragraph – 2]; “A graph cut-based optimization method is run to minimizing the energy function that prefers connected regions having the same label. By repeating the two-step procedure until it converges, the enhanced coarse masks are obtained”.
Summary of Citations (Nicolas)
[Section – 3.4, Paragraph – 1]; “As shown in Fig. 6, the distinction between background and a moving object is implemented using a binary mask that indicates the pixels corresponding to the moving object. In our algorithm, a coarse mask is computed by splitting the intensity frame into subblocks (8 × 8 pixels) ... The mask derived from event statistics and the mask derived from the error between the background and the current frame are combined using a region growing algorithm. Starting from a seed (the event-derived mask), the algorithm iteratively integrates neighboring blocks if they are marked in the error mask. This strategy makes it possible to minimize the amount of false positive (MSE outliers due to strong turbulence incorrectly classified as the moving object) and false negative (flat zones of the moving object incorrectly classified as background)”.
[Page – 17, Paragraph – 1]; “it helps in the segmentation of the moving object without being limited by the comparison of two integrated frames such as in an optical flow-based algorithm. Second, it improves the reconstruction of the moving object appearance and the refinement of its boundaries”.
Summary of Citations (Xuelian)
[Page – 11, Paragraph – 2]; “Given five consecutive frames {It, It+1,It+2,It+3,It+4} and labelled ground-truth gtt, we first estimate forward and backward optical flow fields between frame It and It+n,n ∈ [1,4]”.
[Page – 11, Paragraph – 5]; “Bidirectional Consistency Check. To identify valid masks, we adopt forward-backward consistency check to eliminate inconsistent regions”.
Summary of Citations (Pan)
Paragraph [0001]; “The present invention relates to methods of video stabilization and artefacts removal for turbulence effect compensation in long distance imaging”.
Paragraph [0014]; “there is provided a method of correcting for a turbulence effect in a video comprising a plurality of frames”.
Paragraph [0063]; “The method 200 then proceeds from sub-process 240 to decision step 250 . Step 250 determines whether there are more raw turbulence video frames of the video to be processed. If there are more turbulence video frames to be processed (YES), the method 200 proceeds from step 250 to step 210 . Therefore, sub-processes 210 to 240 are performed for each turbulence video frame”.
Regarding claim 3, Nicolas in the combination discloses the method of claim 1, wherein the method further comprises: generating a coarse map from the coarse masks generated for each moving object; and incorporating a grouping loss function to rectify one or more errors in the coarse map (Nicolas in [Section – 3.4, Paragraph – 1] discloses, “a coarse mask is computed by splitting the intensity frame into subblocks (8 × 8 pixels). ... However, creating a mask based solely on event statistics would only give information about the edges of the moving object ..., we also compare the last background estimate with the current frame and compute the block-wise mean-squared error (MSE) between the two images”).
Summary of Citations (Nicolas)
[Section – 3.4, Paragraph – 1]; “a coarse mask is computed by splitting the intensity frame into subblocks (8 × 8 pixels). ... However, creating a mask based solely on event statistics would only give information about the edges of the moving object ..., we also compare the last background estimate with the current frame and compute the block-wise mean-squared error (MSE) between the two images”.
Regarding claim 10, Nicolas in the combination discloses the method of claim 1, wherein the method further comprises: utilizing a detect-then-grow function to segment moving objects, wherein the detect- then-grow function includes at least: selecting seedling pixels from motion feature maps; and expanding segmentation masks using a region-growing algorithm (Nicolas in [Section – 3.4, Paragraph – 1] discloses, “The mask derived from event statistics and the mask derived from the error between the background and the current frame are combined using a region growing algorithm. Starting from a seed (the event-derived mask), the algorithm iteratively integrates neighboring blocks if they are marked in the error mask. This strategy makes it possible to minimize the amount of false positive (MSE outliers due to strong turbulence incorrectly classified as the moving object”).
Summary of Citations (Nicolas)
[Section – 3.4, Paragraph – 1]; “The mask derived from event statistics and the mask derived from the error between the background and the current frame are combined using a region growing algorithm. Starting from a seed (the event-derived mask), the algorithm iteratively integrates neighboring blocks if they are marked in the error mask. This strategy makes it possible to minimize the amount of false positive (MSE outliers due to strong turbulence incorrectly classified as the moving object”.
Regarding claim 11, apparatus claim 11 corresponds to method claim 1. Therefore, the
rejection analysis and motivation to combine claim 1 is applicable to claim 11.
Regarding claim 13, apparatus claim 13 corresponds to method claim 3. Therefore, the
rejection analysis and motivation to combine claim 3 is applicable to claim 13.
Regarding claim 20, claim 20 is a computer readable storage media corresponds to method claim 1. Therefore, the rejection analysis and motivation to combine claim 1 is applicable to claim 20.
Claims 2 and 12 are rejected under 35 U.S.C 103 as being unpatentable over Pan in view of Nicolas, Dias, Longlong and Xuelian further in view of Van Hook et al, US Patent Application Publication No. US-20230046545-A1 (hereinafter Van).
Regarding claim 2, Pan in the combination discloses the method of claim 1.
Pan, Nicolas, Dias, Longlong and Xuelian in the combination don’t disclose the following limitation as further recited in the claim.
Van discloses capturing long-range imaging video data at a time and place exhibiting environmental atmospheric turbulence which satisfies a threshold atmospheric turbulence condition (Van in [0004] discloses, “In wide field-of-view video acquisition over long ranges, atmospheric optical turbulence undesirably imparts geometric distortion and blurring to images”. Van in [0048] disclose about threshold, “Table 2 of FIG. 2B provides parameters for C.sub.n.sup.2, theoretical r.sub.0, theoreticalD/r.sub.0, isoplanatic angle, RMS Z-tilt, and residual RMS tilt for turbulence degradation Levels 1-6. One of skill will recognize other values are suitable. The present invention has been found to be beneficially in very light to very heavy turbulence”).
It would have been obvious to one of ordinary skill in art before the effective filling date of the claimed invention to integrate the technique of Van into the system of Pan view of Nicolas, Dias, Longlong and Xuelian because it would allow the system improve reliability and quality for real world long distance video affected by strong air.
Summary of Citations (Van)
Paragraph [0004]; “In wide field-of-view video acquisition over long ranges, atmospheric optical turbulence undesirably imparts geometric distortion and blurring to images”.
Paragraph [0048]; “Table 2 of FIG. 2B provides parameters for C.sub.n.sup.2, theoretical r.sub.0, theoreticalD/r.sub.0, isoplanatic angle, RMS Z-tilt, and residual RMS tilt for turbulence degradation Levels 1-6. One of skill will recognize other values are suitable. The present invention has been found to be beneficially in very light to very heavy turbulence”.
Regarding claim 12, apparatus claim 12 corresponds to method claim 2. Therefore, the
rejection analysis and motivation to combine claim 2 is applicable to claim 12.
Claims 4, 5, 6, 14, 15, and 16 are rejected under 35 U.S.C 103 as being unpatentable over Pan in view of Nicolas, Dias, Longlong and Xuelian further in view of Lee US Patent Application Publication No. US-20250029386-A1 (hereinafter Lee).
Regarding claim 4, Pan in the combination discloses the method of claim 1.
Pan, Nicolas, Dias, Longlong and Xuelian in the combination don’t disclose the following limitation as further recited in the claim.
Lee discloses the method further comprises: grouping nearby pixels within the coarse masks across multiple frames of the video frame sequence together to reduce gaps between the nearby pixels (Lee in [0046] discloses, “Any matches of overlapping segmented objects (e.g., a match of a segmented object in the set of segmented objects Dt with a match of an estimated masked object ... By taking the union of the overlapping/matched segmented objects, the association module 320 merges matched segmented objects in the set Dt with estimated masks in the estimated masks, creating a single estimated masked object from two masked/segmented objects”. Overlapping/matched segmented objects are merged equates to grouping pixels. Lee in [0003] discloses about coarse mask. Furthermore, Lee in [0023] discloses about multiple frames. Lee in [0055] discloses about correct segmentation error (reduce gaps between the nearby pixels).
It would have been obvious to one of ordinary skill in art before the effective filling date of the claimed invention to integrate the technique of Lee into the system of Pan view of Nicolas, Dias, Longlong and Xuelian because it would allow the system to correct the disparities and produce more unified and precise segments.
Summary of Citations (Lee)
Paragraph [0003]; “The output of both the image segmentation sub-module and the mask propagation sub-module are compared to obtain a coarse estimation of the masked objects in a frame.”
Paragraph [0023]; “the mask propagation sub-module 110 temporally associates object segments using any mask propagation model. Specifically, the mask propagation sub-module 110 performs pixel-level tracking to track the temporal coherence of object motion across frames. In other words, the mask propagation sub-module 110 tracks segmented objects of each frame using memory of segmented objects in previous frames”
Paragraph [0046]; “Any matches of overlapping segmented objects (e.g., a match of a segmented object in the set of segmented objects Dt with a match of an estimated masked object ... By taking the union of the overlapping/matched segmented objects, the association module 320 merges matched segmented objects in the set Dt with estimated masks in the estimated masks, creating a single estimated masked object from two masked/segmented objects”.
Paragraph [0055]; “At numeral 11, the mask propagation sub-module 110 self-propagates the coherent mask of the current frame to generate a refined mask of the current frame 302. Refining the coherent mask by propagating the coherent mask through the mask propagation sub-module corrects for segmentation errors as a result of the merged segmented objects”.
Regarding claim 5, Pan in the combination discloses the method of claim 1.
Pan, Nicolas, Dias, Longlong and Xuelian in the combination don’t disclose the following limitation as further recited in the claim.
Lee discloses the method further comprises: grouping nearby pixels within the coarse masks across multiple frames of the video frame sequence together to eliminate one or more segment errors between the multiple frames of the video frame sequence (Lee in [0046] discloses about grouping nearby pixels, [0023] discloses about multiple frames (across frames) and [0003] discloses about coarse mask. Lastly, Lee in [0055] discloses about correcting segmentation errors, “the mask propagation sub-module 110 self-propagates the coherent mask of the current frame to generate a refined mask of the current frame 302. Refining the coherent mask by propagating the coherent mask through the mask propagation sub-module corrects for segmentation errors as a result of the merged segmented objects”.
Summary of Citations (Lee)
Paragraph [0003]; “The output of both the image segmentation sub-module and the mask propagation sub-module are compared to obtain a coarse estimation of the masked objects in a frame.”
Paragraph [0023]; “the mask propagation sub-module 110 temporally associates object segments using any mask propagation model. Specifically, the mask propagation sub-module 110 performs pixel-level tracking to track the temporal coherence of object motion across frames. In other words, the mask propagation sub-module 110 tracks segmented objects of each frame using memory of segmented objects in previous frames”
Paragraph [0046]; “Any matches of overlapping segmented objects (e.g., a match of a segmented object in the set of segmented objects Dt with a match of an estimated masked object ... By taking the union of the overlapping/matched segmented objects, the association module 320 merges matched segmented objects in the set Dt with estimated masks in the estimated masks, creating a single estimated masked object from two masked/segmented objects”.
Paragraph [0055]; “At numeral 11, the mask propagation sub-module 110 self-propagates the coherent mask of the current frame to generate a refined mask of the current frame 302. Refining the coherent mask by propagating the coherent mask through the mask propagation sub-module corrects for segmentation errors as a result of the merged segmented objects”.
Regarding claim 6, Pan in the combination discloses the method of claim 1.
Pan, Nicolas, Dias, Longlong and Xuelian in the combination don’t disclose the following limitation as further recited in the claim.
Lee discloses the method further comprises: utilizing a refine-net, optimizing the coarse masks generated for each moving object for spatial-context and consistency across consecutive frames within the video frame sequence to generate the refined masks (Lee in [0056] discloses, “Then, the mask propagation sub-module 110 propagates the estimated segmented representation of the current frame (e.g., [Image Omitted] ), which is based on the coherent mask and features of previous frames, to the current frame 302. As a result, the mask propagation sub-module 110 generates a refined mask Mt of the current frame 302 based on the spatial and/or positional information learned from the segmented representations of previous ht-1 (e.g., a feature map of a previous frame)”).
Summary of Citations (Lee)
Paragraph [0056]; “Then, the mask propagation sub-module 110 propagates the estimated segmented representation of the current frame (e.g., [Image Omitted] ), which is based on the coherent mask and features of previous frames, to the current frame 302. As a result, the mask propagation sub-module 110 generates a refined mask Mt of the current frame 302 based on the spatial and/or positional information learned from the segmented representations of previous ht-1 (e.g., a feature map of a previous frame)”.
Regarding claim 14, apparatus claim 14 corresponds to method claim 4. Therefore, the
rejection analysis and motivation to combine claim 4 is applicable to claim 14.
Regarding claim 15, apparatus claim 15 corresponds to method claim 5. Therefore, the
rejection analysis and motivation to combine claim 5 is applicable to claim 15.
Regarding claim 16, apparatus claim 16 corresponds to method claim 6. Therefore, the
rejection analysis and motivation to combine claim 6 is applicable to claim 16.
Claims 7 and 17 are rejected under 35 U.S.C 103 as being unpatentable over Pan in view of Nicolas, Dias, Longlong and Xuelian further in view of Lv et al, US-12632974-B1 (hereinafter Lv).
Regarding claim 7, Pan in the combination discloses the method of claim 1.
Pan, Nicolas, Dias, Longlong and Xuelian in the combination don’t disclose the following limitation as further recited in the claim.
Lv discloses generating motion feature maps using an epipolar geometry-based consistency check to distinguish between rigid object motion and turbulence-induced or camera-induced motion (Lv in [Column – 1, Line 33 – 43] discloses, “In at least one embodiment, deep learning and data-driven approaches can be used for both static and dynamic depth estimation, as well as estimation of three-dimensional object motion ... higher accuracy and better generalization of performance can be achieved by enforcing geometric constraints derived from epipolar geometry between sequential images captured by a camera”. Furthermore, Lv in [Column – 6, Line 55 – 58] discloses, “In at least one embodiment, moving objects are assumed to be independent of each other, so that rigid motion of each object can be estimated independently”).
It would have been obvious to one of ordinary skill in art before the effective filling date of the claimed invention to integrate the technique of Lv into the system of Pan view of Nicolas, Dias, Longlong and Xuelian because it would make fewer false detections and more accurate moving object segmentation in distorted video.
Summary of Citations (Lv)
[Column – 1, Line 33 – 43]; “In at least one embodiment, deep learning and data-driven approaches can be used for both static and dynamic depth estimation, as well as estimation of three-dimensional object motion, using images captured by one or more monocular cameras. In at least one embodiment, learned priors can enable depth estimation from a single image, a task that can be ill-posed from geometric constraints alone. In at least one embodiment, higher accuracy and better generalization of performance can be achieved by enforcing geometric constraints derived from epipolar geometry between sequential images captured by a camera”.
[Column – 6, Line 55 – 58]; “In at least one embodiment, moving objects are assumed to be independent of each other, so that rigid motion of each object can be estimated independently”.
Regarding claim 17, apparatus claim 17 corresponds to method claim 7. Therefore, the
rejection analysis and motivation to combine claim 7 is applicable to claim 17.
Claims 8 and 18 are rejected under 35 U.S.C 103 as being unpatentable over Pan in view of Nicolas, Dias, Longlong and Xuelian further in view of Li US Patent Publication No. US-20200250832-A1 (hereinafter Li).
Regarding claim 8, Pan in the combination discloses the method of claim 1.
Pan, Nicolas, Dias, Longlong and Xuelian in the combination don’t disclose the following limitation as further recited in the claim.
Li discloses stabilizing optical flow estimations by averaging bidirectional optical flow sequences within a short temporal interval to reduce errors introduced by atmospheric turbulence while preserving features of rigid motion (Li in [0017] discloses about consecutive frame and [0044] successive frame which equates to bidirectional optical flow. Furthermore, Li in [0044] discloses about averaging temporal interval, “he identified plurality of merged objects and the identified plurality of split objects within the successive frames of the image sequence may be done based on different parameters specified in temporal information for a location, a velocity, an acceleration, a size, and an average value of optical flow for the detected plurality of moving regions”).
It would have been obvious to one of ordinary skill in art before the effective filling date of the claimed invention to integrate the technique of Li into the system of Pan in view of Nicolas, Dias, Longlong and Xuelian because the averaging step would help to suppress the random elements and preserve the motion pattern that persists over neighboring frames.
Summary of Citations (Li)
Paragraph [0017]; “the circuitry may be further configured to detect a plurality of moving regions in an image sequence based on an optical flow map of at least two consecutive frames of the image sequence. The optical flow map may be further generated, based on a displacement of a plurality of pixels in the at least two consecutive frames of the image sequence”.
Paragraph [0044]; “he identified plurality of merged objects and the identified plurality of split objects within the successive frames of the image sequence may be done based on different parameters specified in temporal information for a location, a velocity, an acceleration, a size, and an average value of optical flow for the detected plurality of moving regions.”.
Regarding claim 18, apparatus claim 18 corresponds to method claim 8. Therefore, the
rejection analysis and motivation to combine claim 8 is applicable to claim 18.
Claims 9 and 19 are rejected under 35 U.S.C 103 as being unpatentable over Pan in view of Nicolas, Dias, Longlong and Xuelian further in view of Ma Patent Application Publication No. CN-113111973-A (hereinafter Ma).
Regarding claim 9, Pan in the combination discloses the method of claim 1.
Pan, Nicolas, Dias, Longlong and Xuelian in the combination don’t disclose the following limitation as further recited in the claim.
Ma discloses distinguishing moving objects from a static background using a Sampson distance map computed based on fundamental matrices derived from a stabilized optical flow that quantifies geometric consistency errors associated with the moving objects (Ma in [Page – 6, Paragraph – 3 & 4] discloses, “wherein the second distance represents a first order approximation of the geometric distance; in the embodiment of the invention, based on the constraint of the basic matrix to primarily distinguish the static point and the dynamic point ... the second distance, namely the Sampson distance is a first order approximation of the set distance, describing the distance between the image point and its corresponding pole line in another frame image; it can improve the robustness of the algorithm”).
It would have been obvious to one of ordinary skill in art before the effective filling date of the claimed invention to integrate the technique of Ma into the system of Pan view of Nicolas, Dias, Longlong and Xuelian because it would the system to output more accurate foreground or background separation.
Summary of Citations (Ma)
[Page – 6, Paragraph – 3 & 4]; “wherein the second distance represents a first order approximation of the geometric distance; in the embodiment of the invention, based on the constraint of the basic matrix to primarily distinguish the static point and the dynamic point ... the second distance, namely the Sampson distance is a first order approximation of the set distance, describing the distance between the image point and its corresponding pole line in another frame image; it can improve the robustness of the algorithm”.
Regarding claim 19, apparatus claim 19 corresponds to method claim 9. Therefore, the
rejection analysis and motivation to combine claim 9 is applicable to claim 19.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner
should be directed to ZAID MUHAMMAD SALEH whose telephone number is (703)756-1684.
The examiner can normally be reached M-F 8 am - 5 pm ET. 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, Vu Le can be
reached on (571)272-7332. 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.
/ZAID MUHAMMAD SALEH/
Examiner, Art Unit 2668
09/17/2026
/VU LE/Supervisory Patent Examiner, Art Unit 2668
Prosecution Insights