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 .
Priority
Receipt is acknowledged that application claims priority to foreign application with application number CHINA 202311723077.7 dated 12/14/2023.
Copies of certified papers required by 37 CFR 1.55 have been received. Priority is acknowledged under 35 USC 119(e) and 37 CFR 1.78.
Information Disclosure Statement
The IDS(s) dated 1/14/2025 has/have been considered and placed in the application file.
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, 6, 9, 14, and 17 is/are rejected under 35 U.S.C. 103 as obvious over Senkal et al (US 20220206566 A1, hereafter referred to as Senkal) in view of Li et al (Li, Y., Wang, G., Ji, X., Xiang, Y., & Fox, D. (2018). Deepim: Deep iterative matching for 6d pose estimation. In Proceedings of the European conference on computer vision (ECCV) (pp. 683-698), hereafter referred to as Li).
Claim 1
Regarding Claim 1, Senkal teaches “A method for determining a position, comprising:
inputting a historical time queue (Senkal in ¶84 discloses timestamped input sequences with optional position features from prior determinations) and posture change information of a target object at a target point of time (Senkal in ¶¶82-84 discloses timestamped linear/angular acceleration, optional orientation/velocity features, and tracking for the last time step) to a position estimation model (Senkal in ¶¶82-88 discloses a temporal CNN producing tracking position/pose information) to obtain an initial predicted position (Senkal in ¶¶82-88 discloses a temporal CNN producing tracking position/pose information; ¶¶104-106 discloses reconstructing an updated position from predicted motion and runtime camera correction),
wherein the historical time queue is used for storing historical position information of the target object at latest n historical points of time prior to the target point of time (Senkal in ¶84 discloses timestamped input sequences with optional position features from prior determinations; ¶¶55-58 discloses successive recent time windows and prior position conditioning to narrow possible positions), and n is a preset positive integer not less than 2 (Senkal in ¶84 discloses a preset 128 step input window) … .”
Senkal does not explicitly teach all of “… performing at least two iterative stages on the initial predicted position to obtain position information of the target object at the target point of time, wherein a positioning accuracy of any iterative stage is higher than a positioning accuracy of a previous iterative stage.”
However, Li teaches “… performing at least two iterative stages on the initial predicted position to obtain position information of the target object at the target point of time (Li in §4.1, p. 9 discloses two testing iteration; §3.5, p. 8 discloses applying each predicted correction and using the updated pose for the next pass),
wherein a positioning accuracy of any iterative stage is higher than a positioning accuracy of a previous iterative stage (Li in pp. 2, 5-8, 10, §§3.2-3.3, 3.5, 4.3, Table 1 discloses progressively more accurate pose estimates, separate translation refinement, iterative training, and better first to second test stage performance for a two iteration trained network).”
Li is analogous art because Li, like Senkal, is in the field of learned object localization for virtual reality and is reasonably pertinent to the problem of correcting an inaccurate initial target position estimate.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the current position tracking and camera correction processing of Senkal with the two stage, iteratively trained rendered/observed image pose refinement of Li because Li teaches that iterative correction and re-rendering provides progressively closer image matching and more accurate pose estimates (pp.2, 8, 10, §§3.5, 4.3), and one of ordinary skill would have recognized that incorporating this feature into the method of Senkal would correct the initial current pose, re-render the object at that corrected pose, and correct the pose again before outputting the refined target position, with a reasonable expectation of success because Senkal supplies current position/orientation estimates and runtime tracking images, Li accepts an externally supplied initial pose through its image/model interfaces, and the registered translation conversion and iterative training preserve those interfaces while supporting first to second stage improvement (Senkal, ¶¶57, 88, 104, 106 -107; Li, §§3.1-3.3, 3.5, pp.4-8, 10-12, Tables 1, 3).
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Claim 6
Regarding Claim 6, Senkal in view of Li teaches The method according to claim 1, wherein the performing at least two iterative stages according to the initial predicted position to obtain position information of the target object at the target point of time comprises:
determining a relative position of the target object at the target point of time to a previous historical point of time (Senkal in ¶¶55, 82, 88, 104 discloses tracking temporal motion and updating a previously determined position), and using the relative position as the position information of the target object at the target point of time (Senkal in ¶88 discloses tracking motion/position change output); or
determining a relative position of the target object at the target point of time to a previous historical point of time (Senkal in ¶¶55, 82, 88, 104 discloses tracking temporal motion and updating a previously determined position), and determining the position information of the target object at the target point of time according to the relative position and historical position information of the target object at the previous historical point of time (Senkal in ¶104 discloses updating an initial position using predicted temporal motion; Li in pp. 6-8, §§3.3, 3.5, Eq. 2 discloses translation correction and iterative pose updating).
Claim 9
Regarding Claim 9, Senkal teaches “An electronic device, comprising:
at least one memory and at least one processor (Senkal in ¶¶132-133 discloses processor and memory);
wherein the at least one memory is configured to store a program code, and the at least one processor is configured to execute the program code stored on the at least one memory and cause the electronic device to:
input a historical time queue (Senkal in ¶84 discloses timestamped input sequences with optional position features from prior determinations) and posture change information of a target object at a target point of time (Senkal in ¶¶82-84 discloses timestamped linear/angular acceleration, optional orientation/velocity features, and tracking for the last time step) to a position estimation model (Senkal in ¶¶82-88 discloses a temporal CNN producing tracking position/pose information) to obtain an initial predicted position (Senkal in ¶¶82-88 discloses a temporal CNN producing tracking position/pose information; ¶¶104-106 discloses reconstructing an updated position from predicted motion and runtime camera correction),
wherein the historical time queue is used for storing historical position information of the target object at latest n historical points of time prior to the target point of time (Senkal in ¶84 discloses timestamped input sequences with optional position features from prior determinations; ¶¶55-58 discloses successive recent time windows and prior position conditioning to narrow possible positions), and n is a preset positive integer not less than 2 (Senkal in ¶84 discloses a preset 128 step input window) … .”
Senkal does not explicitly teach all of “… perform at least two iterative stages on the initial predicted position to obtain position information of the target object at the target point of time, wherein a positioning accuracy of any iterative stage is higher than a positioning accuracy of a previous iterative stage.”
However, Li teaches “… perform at least two iterative stages on the initial predicted position to obtain position information of the target object at the target point of time (Li in §4.1, p. 9 discloses two testing iteration; §3.5, p. 8 discloses applying each predicted correction and using the updated pose for the next pass),
wherein a positioning accuracy of any iterative stage is higher than a positioning accuracy of a previous iterative stage (Li in pp. 2, 5-8, 10, §§3.2-3.3, 3.5, 4.3, Table 1 discloses progressively more accurate pose estimates, separate translation refinement, iterative training, and better first to second test stage performance for a two iteration trained network).”
Li is analogous art because Li, like Senkal, is in the field of learned object localization for virtual reality and is reasonably pertinent to the problem of correcting an inaccurate initial target position estimate.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the current position tracking and camera correction processing of Senkal with the two stage, iteratively trained rendered/observed image pose refinement of Li because Li teaches that iterative correction and re-rendering provides progressively closer image matching and more accurate pose estimates (pp.2, 8, 10, §§3.5, 4.3), and one of ordinary skill would have recognized that incorporating this feature into the method of Senkal would correct the initial current pose, re-render the object at that corrected pose, and correct the pose again before outputting the refined target position, with a reasonable expectation of success because Senkal supplies current position/orientation estimates and runtime tracking images, Li accepts an externally supplied initial pose through its image/model interfaces, and the registered translation conversion and iterative training preserve those interfaces while supporting first to second stage improvement (Senkal, ¶¶57, 88, 104, 106 -107; Li, §§3.1-3.3, 3.5, pp.4-8, 10-12, Tables 1, 3).
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Claim 14
Regarding Claim 14, Senkal in view of Li teaches The electronic device according to claim 9, wherein the electronic device is further caused to:
determine a relative position of the target object at the target point of time to a previous historical point of time (Senkal in ¶¶55, 82, 88, 104 discloses tracking temporal motion and updating a previously determined position), and using the relative position as the position information of the target object at the target point of time (Senkal in ¶88 discloses tracking motion/position change output); or
determine a relative position of the target object at the target point of time to a previous historical point of time (Senkal in ¶¶55, 82, 88, 104 discloses tracking temporal motion and updating a previously determined position), and determining the position information of the target object at the target point of time according to the relative position and historical position information of the target object at the previous historical point of time (Senkal in ¶104 discloses updating an initial position using predicted temporal motion; Li in pp. 6-8, §§3.3, 3.5, Eq. 2 discloses translation correction and iterative pose updating).
Claim 17
Regarding Claim 17, Senkal teaches “A computer-readable storage medium, configured to store a program code which, when executed by a processor, causes the processor to:
input a historical time queue (Senkal in ¶84 discloses timestamped input sequences with optional position features from prior determinations) and posture change information of a target object at a target point of time (Senkal in ¶¶82-84 discloses timestamped linear/angular acceleration, optional orientation/velocity features, and tracking for the last time step) to a position estimation model (Senkal in ¶¶82-88 discloses a temporal CNN producing tracking position/pose information) to obtain an initial predicted position (Senkal in ¶¶82-88 discloses a temporal CNN producing tracking position/pose information; ¶¶104-106 discloses reconstructing an updated position from predicted motion and runtime camera correction),
wherein the historical time queue is used for storing historical position information of the target object at latest n historical points of time prior to the target point of time (Senkal in ¶84 discloses timestamped input sequences with optional position features from prior determinations; ¶¶55-58 discloses successive recent time windows and prior position conditioning to narrow possible positions), and n is a preset positive integer not less than 2 (Senkal in ¶84 discloses a preset 128 step input window) … .”
Senkal does not explicitly teach all of “… perform at least two iterative stages on the initial predicted position to obtain position information of the target object at the target point of time, wherein a positioning accuracy of any iterative stage is higher than a positioning accuracy of a previous iterative stage.”
However, Li teaches “… perform at least two iterative stages on the initial predicted position to obtain position information of the target object at the target point of time (Li in §4.1, p. 9 discloses two testing iteration; §3.5, p. 8 discloses applying each predicted correction and using the updated pose for the next pass),
wherein a positioning accuracy of any iterative stage is higher than a positioning accuracy of a previous iterative stage (Li in pp. 2, 5-8, 10, §§3.2-3.3, 3.5, 4.3, Table 1 discloses progressively more accurate pose estimates, separate translation refinement, iterative training, and better first to second test stage performance for a two iteration trained network).”
Li is analogous art because Li, like Senkal, is in the field of learned object localization for virtual reality and is reasonably pertinent to the problem of correcting an inaccurate initial target position estimate.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the current position tracking and camera correction processing of Senkal with the two stage, iteratively trained rendered/observed image pose refinement of Li because Li teaches that iterative correction and re-rendering provides progressively closer image matching and more accurate pose estimates (pp.2, 8, 10, §§3.5, 4.3), and one of ordinary skill would have recognized that incorporating this feature into the method of Senkal would correct the initial current pose, re-render the object at that corrected pose, and correct the pose again before outputting the refined target position, with a reasonable expectation of success because Senkal supplies current position/orientation estimates and runtime tracking images, Li accepts an externally supplied initial pose through its image/model interfaces, and the registered translation conversion and iterative training preserve those interfaces while supporting first to second stage improvement (Senkal, ¶¶57, 88, 104, 106 -107; Li, §§3.1-3.3, 3.5, pp.4-8, 10-12, Tables 1, 3).
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Claim(s) 7 and 15 is/are rejected under 35 U.S.C. 103 as obvious over Senkal et al (US 20220206566 A1, hereafter referred to as Senkal) and Li et al (Li, Y., Wang, G., Ji, X., Xiang, Y., & Fox, D. (2018). Deepim: Deep iterative matching for 6d pose estimation. In Proceedings of the European conference on computer vision (ECCV) (pp. 683-698), hereafter referred to as Li), further in view of Bai et al (Bai, S., Kolter, J. Z., & Koltun, V. (2018). An empirical evaluation of generic convolutional and recurrent networks for sequence modeling. arXiv preprint arXiv:1803.01271, hereafter referred to as Bai).
Claim 7
Regarding Claim 7, Senkal in view of Li teaches “The method according to claim 1 …”
Senkal in view of Li does not explicitly teach all of “… wherein: the position estimation model is a dilated convolution neural network model, and a dilatation coefficient of the dilated convolution neural network model is not less than 2.”
However, Bai teaches “… wherein: the position estimation model is a dilated convolution neural network model (Bai in pp. 3-4, §§3.1-3.3, Fig. 1(a) discloses casual dilated temporal convolution), and a dilatation coefficient of the dilated convolution neural network model is not less than 2 (Bai in p.4, §3.3, Fig. 1(a) discloses dilation factors d=1,2,4 and increasing dilation with depth).”
Bai is analogous art because Bai, like Senkal in view of Li, is in the field of neural prediction and is reasonably pertinent to the problem of obtaining adequate historical coverage for the current position prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the temporal CNN feature extractor of Senkal in view of Li with the casual dilated temporal modules of Bai because Bai teaches that increased dilation provides a wider receptive field and controllable effective history without the extremely deep network or very large filters of basic casual convolution (§§3.2-3.3, 3.5, pp. 3-5), and one of ordinary skill would have recognized that incorporating this feature into the method of Senkal in view of Li would generate a current time position estimate informed by the selected longer historical window before the retained refinement, with a reasonable expectation of success because both architectures process temporal feature tensors, Senkal allows alternative neural architectures and accepts CNN features at its dense/output head, and task specific training with a window matched receptive field preserves the estimator interfaces.
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Claim 15
Regarding Claim 15, Senkal in view of Li teaches “The electronic device according to claim 9 …”
Senkal in view of Li does not explicitly teach all of “… wherein: the position estimation model is a dilated convolution neural network model, and a dilatation coefficient of the dilated convolution neural network model is not less than 2.”
However, Bai teaches “… wherein: the position estimation model is a dilated convolution neural network model (Bai in pp. 3-4, §§3.1-3.3, Fig. 1(a) discloses casual dilated temporal convolution), and a dilatation coefficient of the dilated convolution neural network model is not less than 2 (Bai in p.4, §3.3, Fig. 1(a) discloses dilation factors d=1,2,4 and increasing dilation with depth).”
Bai is analogous art because Bai, like Senkal in view of Li, is in the field of neural prediction and is reasonably pertinent to the problem of obtaining adequate historical coverage for the current position prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the temporal CNN feature extractor of Senkal in view of Li with the casual dilated temporal modules of Bai because Bai teaches that increased dilation provides a wider receptive field and controllable effective history without the extremely deep network or very large filters of basic casual convolution (§§3.2-3.3, 3.5, pp. 3-5), and one of ordinary skill would have recognized that incorporating this feature into the method of Senkal in view of Li would generate a current time position estimate informed by the selected longer historical window before the retained refinement, with a reasonable expectation of success because both architectures process temporal feature tensors, Senkal allows alternative neural architectures and accepts CNN features at its dense/output head, and task specific training with a window matched receptive field preserves the estimator interfaces.
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Claim(s) 8 and 16 is/are rejected under 35 U.S.C. 103 as obvious over Senkal et al (US 20220206566 A1, hereafter referred to as Senkal), Li et al (Li, Y., Wang, G., Ji, X., Xiang, Y., & Fox, D. (2018). Deepim: Deep iterative matching for 6d pose estimation. In Proceedings of the European conference on computer vision (ECCV) (pp. 683-698), hereafter referred to as Li), and Bai et al (Bai, S., Kolter, J. Z., & Koltun, V. (2018). An empirical evaluation of generic convolutional and recurrent networks for sequence modeling. arXiv preprint arXiv:1803.01271, hereafter referred to as Bai), further in view of Gal et al (Gal, Y., & Ghahramani, Z. (2016, June). Dropout as a Bayesian approximation: Representing model uncertainty in deep learning. In international conference on machine learning (pp. 1050-1059). PMLR., hereafter referred to as Gal).
Claim 8
Regarding Claim 8, Senkal and Li, further in view of Bai teaches “The method according to claim 7, wherein the following operations are performed at each convolutional layer of the position estimation model:
performing preset processing on layer input data of a current convolutional layer twice to obtain layer output data (Bai in §3.4, p. 4, Fig. 1(b) discloses two sequential weight normalized dilated convolution/ReLU/dropout operations producing the residual branch transform), or performing preset processing on layer input data at least once and then combining the processed data with data from 1×1 convolution on the layer input data to obtain layer output data (Bai in §3.4, p. 4, Fig. 1(b) discloses two processing sequences and addition of the input passed through a 1x1 projection when feature widths differ),
wherein layer input data of a first layer of the position estimation model is the historical time queue and the posture change information (Senkal in ¶84-85 discloses timestamped motion features, optional prior position inputs, and optional averaging/decimation of consecutive samples) and the layer output data of the current convolutional layer is layer input data of next convolutional layer (Bali in §§3.2, 3.4, pp. 3-4, Fig. 1 discloses casual stacked layers with residual module outputs feeding subsequent modules); and
wherein the preset processing comprises: performing weight parameter normalized dilated convolution processing on the layer input data and then performing non-linear processing using an activation function (Bali in §3.4, p. 4, Fig 1(b) discloses weight normalization applied to the dilated convolutional filters, followed by ReLU activation in each of the two processing sequences) … .”
Senkal and Li, further in view of Bai does not explicitly teach all of “ … performing processing by a discarding unit.”
However, Gal teaches “ … performing processing by a discarding unit (Gal in §§3-4, pp. 2-4, Eq. 6 discloses stochastic inference through dropout trained networks and output averaging).”
Gal is analogous art because Gal, like Senkal and Li, further in view of Bai, is in the field of neural regression models and is reasonably pertinent to the problem of deriving a predictive mean and uncertainty from a dropout equipped position estimator.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the inference processing of Senkal and Li, further in view of Bai with the stochastic dropout forward passes and output averaging of Gal because Gal teaches that this processing provides a predictive mean and grounded uncertainty estimates from an existing dropout trained network (§4, p. 4, Eq (6)), and one of ordinary skill would have recognized that incorporating this feature into the temporal estimator of Senkal and Li, further in view of Bai would actively apply the trained dropout operations during position prediction and produce a model averaged initial position for the retained refinement rather than relying exclusively on one weighted average prediction, with a reasonable expectation of success because the model is trained with the same dropout locations and mask distribution, masking preserves tensor dimensions, the projection paths preserve residual addition compatibility, and averaging metric position outputs preserves the initial position interface to Li.
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Claim 16
Regarding Claim 16, Senkal and Li, further in view of Bai teaches “The electronic device according to claim 15, wherein the electronic device is caused to perform the following operations are performed at each convolutional layer of the position estimation model:
performing preset processing on layer input data of a current convolutional layer twice to obtain layer output data (Bai in §3.4, p. 4, Fig. 1(b) discloses two sequential weight normalized dilated convolution/ReLU/dropout operations producing the residual branch transform), or performing preset processing on layer input data at least once and then combining the processed data with data from 1×1 convolution on the layer input data to obtain layer output data (Bai in §3.4, p. 4, Fig. 1(b) discloses two processing sequences and addition of the input passed through a 1x1 projection when feature widths differ),
wherein layer input data of a first layer of the position estimation model is the historical time queue and the posture change information (Senkal in ¶84-85 discloses timestamped motion features, optional prior position inputs, and optional averaging/decimation of consecutive samples) and the layer output data of the current convolutional layer is layer input data of next convolutional layer (Bali in §§3.2, 3.4, pp. 3-4, Fig. 1 discloses casual stacked layers with residual module outputs feeding subsequent modules); and
wherein the preset processing comprises: performing weight parameter normalized dilated convolution processing on the layer input data and then performing non-linear processing using an activation function (Bali in §3.4, p. 4, Fig 1(b) discloses weight normalization applied to the dilated convolutional filters, followed by ReLU activation in each of the two processing sequences) … .”
Senkal and Li, further in view of Bai does not explicitly teach all of “ … performing processing by a discarding unit.”
However, Gal teaches “ … performing processing by a discarding unit (Gal in §§3-4, pp. 2-4, Eq. 6 discloses stochastic inference through dropout trained networks and output averaging).”
Gal is analogous art because Gal, like Senkal and Li, further in view of Bai, is in the field of neural regression models and is reasonably pertinent to the problem of deriving a predictive mean and uncertainty from a dropout equipped position estimator.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the inference processing of Senkal and Li, further in view of Bai with the stochastic dropout forward passes and output averaging of Gal because Gal teaches that this processing provides a predictive mean and grounded uncertainty estimates from an existing dropout trained network (§4, p. 4, Eq (6)), and one of ordinary skill would have recognized that incorporating this feature into the temporal estimator of Senkal and Li, further in view of Bai would actively apply the trained dropout operations during position prediction and produce a model averaged initial position for the retained refinement rather than relying exclusively on one weighted average prediction, with a reasonable expectation of success because the model is trained with the same dropout locations and mask distribution, masking preserves tensor dimensions, the projection paths preserve residual addition compatibility, and averaging metric position outputs preserves the initial position interface to Li.
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Claim(s) 4 and 5 is/are rejected under 35 U.S.C. 103 as obvious over Senkal et al (US 20220206566 A1, hereafter referred to as Senkal) and Li et al (Li, Y., Wang, G., Ji, X., Xiang, Y., & Fox, D. (2018). Deepim: Deep iterative matching for 6d pose estimation. In Proceedings of the European conference on computer vision (ECCV) (pp. 683-698), hereafter referred to as Li), further in view of Chen et al (Chen, K., Zhao, X., Dong, C., Di, Z., & Chen, Z. (2024). Anti-occlusion object tracking algorithm based on filter prediction. Journal of Shanghai Jiaotong University (Science), 29(3), 400-413, hereafter referred to as Chen).
Claim 4
Regarding Claim 4, Senkal in view of Li teaches “The method according to claim 1, before the inputting a historical time queue and posture change information of a target object at a target point of time to a position estimation model to obtain an initial predicted position, further comprising: …”
Senkal in view of Li does not explicitly teach all of “determining whether the target object is located in a shooting blind spot of a camera at n historical points of time in the historical time queue and the target point of time; in response to the target object being located beyond the shooting blind spot at the n historical points of time and the target point of time, using position information of the target object at the target point of time acquired by the camera as the position information of the target object at the target point of time; and in response to the target object being located in the shooting blind spot at at least one point of time selected from the group of the n historical points of time and the target point of time, inputting the historical time queue and the posture change information to the position estimation model to obtain the initial predicted position.”
However, Chen teaches “determining whether the target object is located in a shooting blind spot of a camera at n historical points of time in the historical time queue and the target point of time (Chen in §2.5, p. 406, Fig. 8 discloses immediate occlusion for unknown/not found observations and prediction retention until two consecutive normal frames);
in response to the target object being located beyond the shooting blind spot at the n historical points of time and the target point of time, using position information of the target object at the target point of time acquired by the camera as the position information of the target object at the target point of time (Chen in Figs. 9-10, pp. 406-407 discloses image tracker target center output in its non-occlusion branch); and
in response to the target object being located in the shooting blind spot at at least one point of time selected from the group of the n historical points of time and the target point of time, inputting the historical time queue and the posture change information to the position estimation model to obtain the initial predicted position (Chen in §2.5, p. 406, Figs. 8-9 discloses retaining prediction on the first normal observation after occlusion).”
Chen is analogous art because Chen, like Senkal in view of Li, is in the field of position tracking and is reasonably pertinent to the problem of avoiding premature reliance on a newly recovered camera observation.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify prediction/camera selection control of Senkal in view of Li with the prediction retention control of Chen because Chen teaches that retaining prediction through the first normal observation after a failed observation provides an occlusion handling mechanism that reduces tracking failure on loss and reappearance (§§2.5, 3.3, pp. 405-407, 410-412, Figs. 8-10), and one of ordinary skill would have recognized that incorporating this feature into the method of Senkal in view of Li would use the recorded prior blind observation to retain the temporal prediction on the first image observed recovery frame and then refine that prediction with the recovered current image rather than immediately substituting camera only output, with a reasonable expectation of success because the selected Senkal model already accepts timestamped historical positions and current motion information, the recovery control doesn’t alter the prediction interface, and the current recovered image, known object geometry, registered pose conversion, and retained iterative training supply the refinement prerequisites.
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention.
Claim 5
Regarding Claim 5, Senkal and Li, further in view of Chen teaches The method according to claim 4, wherein:
after determining that the target object is located in the shooting blind spot at at least one point of time selected from the group of the n historical points of time and the target point of time, the initial predicted position is obtained by using the position estimation model (Chen in §2.5, p. 406, Figs. 8-9 discloses prediction after the occlusion/recovery decision; Senkal in ¶82-88 discloses the selected temporal position estimator); and
after determining that the target object is located beyond the shooting blind spot at the n historical points of time and the target point of time, the position estimation model is maintained at or set to a non-operating state (Chen in §2.5, p. 406, Figs. 8-9 discloses prediction after the occlusion/recovery decision; Senkal in ¶82-88 discloses the selected temporal position estimator. The combination is applied to an execution where the target object was in the blind spot at the immediately preceding historical point of time and is visible at the target point of time.).
Allowable Subject Matter
Claims 2-3, 10-13 and 18-20 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JUSTIN P CASCAIS whose telephone number is (703) 756-5576. The examiner can normally be reached Monday-Friday 8:00-4:00.
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, Mr. O'Neal Mistry can be reached on (313) 446-4912. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of 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.
/J.P.C./Examiner, Art Unit 2674
/ONEAL R MISTRY/Supervisory Patent Examiner, Art Unit 2674
Date: 9/10/2026
Prosecution Insights