Prosecution Insights
Last updated: November 20, 2026
Application No. 19/051,516

IMAGE PROCESSING APPARATUS AND IMAGE PROCESSING METHOD

Non-Final OA §101§103§112
Filed
Feb 12, 2025
Priority
Feb 29, 2024 — JP 2024-030637
Examiner
SOFRONIOU, MICHAEL MARIO
Art Unit
2661
Tech Center
2600 — Communications
Assignee
Canon Inc.
OA Round
1 (Non-Final)
100%
Grant Probability
Favorable
1-2
OA Rounds
7m
Est. Remaining
99%
With Interview

Examiner Intelligence

Grants 100% — above average
100%
Career Allowance Rate
5 granted / 5 resolved
+38.0% vs TC avg
Minimal +0% lift
Without
With
+0.0%
Interview Lift
resolved cases with interview
Typical timeline
2y 4m
Avg Prosecution
18 currently pending
Career history
34
Total Applications
across all art units

Statute-Specific Performance

§101
7.7%
-32.3% vs TC avg
§103
44.2%
+4.2% vs TC avg
§102
12.5%
-27.5% vs TC avg
§112
32.7%
-7.3% vs TC avg
Black line = Tech Center average estimate • Based on career data from 5 resolved cases

Office Action

§101 §103 §112
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 Acknowledgment is made of applicant’s claim for foreign priority under 35 U.S.C. 119 (a)-(d). The certified copy of the priority document, JP2024-030637, filed on 02/29/2024, has been received. Claim Objections Claims 1, 14 & 15 objected to because of the following informalities: These claims all recite “wherein for each of areas of the plurality of subjects”. The syntax of the underlined phrase is syntactically confusing. The examiner is unsure if applicant intended to recite something more akin to “wherein for each area of the plurality of subjects” or “wherein for each area of the plurality of areas for each of the plurality of subjects”. Appropriate correction is required. Claim Rejections - 35 USC § 112 The following is a quotation of 35 U.S.C. 112(b): (b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention. The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph: The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention. Claims 1-12, 14 & 15 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention. Claims 1, 14 & 15 recite “based on evaluation values output by the trained-machine learning model”, which first introduces the limitation of “evaluation values”. The claims then further recite “wherein for each of areas of the plurality of subjects, the trained machine-learning model outputs an evaluation value that…”. Here, “an evaluation value” is introduced as if it is defining a new term, however, the antecedent basis for “evaluation value” has already been established earlier in the claim. While the examiner believes these two evaluation values are synonymous, the second introduction of “evaluation value” renders the claim unclear and therefore indefinite. Claims 2-12 are subsequently rejected for being dependent on a rejected claim. Claim 5 recites “the ranking evaluation functions are evaluation functions of the respective rankings”. The underlined limitations lack antecedent basis. “the ranking evaluation function” is first introduced in claim 4, however claim 5 is recited to only depend on claim 3. The examiner is unsure if applicant intended claim 5 to further depend on claim 4, which would alleviate the lack of antecedent basis for “the ranking evaluation function”. Furthermore, “the respective ranking” also lacks antecedent basis. Claim 1 establishes antecedent basis for “the ranking” and “the respective ranking” but nowhere in the chain of dependency of claim 5 is “a respective ranking” every properly defined. Claim Rejections - 35 USC § 101 35 U.S.C. 101 reads as follows: Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title. Claim 1, 12-15 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. Claim 1 is reproduced below with additional annotations: An image processing apparatus, comprising: one or more processors, wherein the one or more processors execute a program stored in a memory and thereby perform a method comprising: obtaining image data of one frame including a plurality of subjects moving in a specific direction; inputting the image data to a trained machine-learning model; based on evaluation values output by the trained machine-learning model, determining rankings of the plurality of subjects in the specific direction; and based on a result of the determination, executing processing for tracking a subject of a specific ranking, and wherein for each of areas of the plurality of subjects, the trained machine-learning model outputs an evaluation value that increases or decreases as the ranking is closer to the specific ranking. Analysis for claim 1 for subject matter eligibility under the Alice/Mayo framework is as follows: Step 1: Evaluating whether the claim belongs to one of the statutory categories. Claim 1 recites a monitoring apparatus which is directed to a machine, which falls under one of the statutory categories of invention. (Step 1: Yes) Step 2A Prong One: Evaluating whether the claim recites a judicial exception (an abstract idea enumerated in the 2019 PEG, a law of nature, or a natural phenomenon). If no exception is recited, the claim is eligible. This concludes the eligibility analysis. If the claim recites an exception, go to Step 2A Prong Two. Independent claim 1 recites a plurality of mental processes, which fall under the category of abstract ideas. Steps f & g of claim 1, under the broadest reasonable interpretation, recite a plurality of processes that could be practically performed in the human mind. (The courts consider a mental process (thinking) that "can be performed in the human mind, or by a human using a pen and paper" to be an abstract idea.) Step f recites a process of utilizing evaluation values provided by a machine-learning model to determine the rankings of subjects in a particular direction. This determination step describes a general process of judgement/evaluation that a person would be able to practically conduct in their mind. A person would be able to use evaluation values provided by an ML model, and use a set of rules to determine which subject would place first, second, third, etc., based on those values. Step g goes on to describe a general process of tracking subjects of a particular ranking based on that determination. A person could readily use image data, in the form of a video, and visually track a subject in the lead. These steps would fall into the “mental process” group of abstract ideas – specifically, observations and judgements. These limitations, interpreted under their broadest reasonable interpretation in consistence with the specification, cover the performance of the limitations in the mind or by generic computer components. See MPEP 2106.04 and the 2019 PEG, with particular attention to Example 46 of Appendix 1. (Step 2A Prong One: Yes). Step 2A Prong Two: Evaluating whether the claim recites additional elements that integrate the exception into a practical application of the exception. This evaluation is performed by (a) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (b) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. Claim 1 also recites additional elements – “one or more processors, wherein the one or more processors execute a program stored in a memory”. The processor and memory are recited at a high level of generality such that they do not amount to more than generic computer system elements. The claim does not direct to a specific improvement in computers in their communication role or provides a specific improvement in the way computers operate. Step d recite a process of obtaining image data of subject moving in a particular direction. The act of obtaining, or otherwise gathering data amounts to no more than insignificant pre-solution activity which is used in conjunction with the abstract ideas recited in steps f & g. Step e recites inputting image data into a trained machine-learning model. This use of a machine-learning model is recited with a high degree of generality, providing little details in the training of the model. At the time of the effective filing date of the present application, using trained models in computers, without significantly more details, is generic claiming of conventional computer techniques, which does not amount to significantly more than the abstract idea. Step h recites additional details regarding the outputs of the machine-learning model, describing that the evaluation values either increase or decrease for a ranking as it approaches a specific ranking. This element merely further defines the relationship between the evaluation values and a ranking and fails to amount to significantly more than the abstract idea. These limitations generally link the use of the abstract idea to the environment of subject tracking and ranking. These additional elements neither integrate the abstract idea into a practical application, nor result in the claim as a whole amounting to significantly more than the abstract idea. (Step 2A Prong Two: No). Step 2B: This part of the eligibility analysis evaluates whether the claim as a whole amounts to significantly more than the recited exception, i.e., whether any additional element, or combination of additional elements, adds an inventive concept to the claim. Claim 1 as a whole does not amount to significantly more than the recited exception, nor does it add an inventive concept to the claim. It merely recites performing a plurality of mental processes with the general implementation of a machine-learning model without significantly more. The act of tracking and ranking subjects moving in a particular direction is well-understood and routine in the art and fails to effectuate the claimed function into a practical application. See MPEP 2016.04(d) & MPEPE 2016.05. (Step 2B: No). Therefore, claim 1 is not eligible. Claim 12 recites the additional limitation of a specific ranking being first place. This element only further limits a particular ranking to track, which a human would readily be able to perform and fails to integrate the claimed invention into a practical application. Therefore, claim 12 is not eligible. Claim 13 recites substantially identical subject matter to claim 1, with the additional limitation of: setting a focus detection area at the detected subject of the specific ranking. This limitation is recited rather broadly, and the act of setting a focus area to be of a subject of a particular ranking can still be considered a mental process. A person could readily focus on a subject of a particular ranking, such as first place, and continuously focus or otherwise observe that subject in the image data. Therefore, claim 13 is not eligible. Claim 14 recites substantially identical subject matter to claim 1, with no additional elements, and follows a similar analysis. Therefore, claim 14 is not eligible. Claim 15 recites an additional element: “A non-transitory computer-readable medium storing a program which, when executed by a computer, cause the computer to perform an image processing method…” non-transitory computer-readable medium and computer are recited at a high level of generality such that they do not amount to more than generic computer system elements. The claim does not direct to a specific improvement in computers in their communication role or provides a specific improvement in the way computers operate. These limitations generally link the use of the abstract idea to the environment of surveillance and pose estimation. These additional elements neither integrate the abstract idea into a practical application, nor result in the claim as a whole amounting to significantly more than the abstract idea. Therefore, claim 15 is not eligible. 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. Claim(s) 1, 3, 8, 9, 12-15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Leung et al. (US 2020/0401793 A1), hereinafter referred to as “Leung”, further in view of Aoyama et al. (US 2022/0005205 A1) hereinafter referred to as “Aoyama”. Regarding claim 1, Leung teach An image processing apparatus (image capture system 110 [¶0037; Fig. 1]), comprising: one or more processors (the camera system module 140 can be implemented by one or more programmable processors [¶0065; Figs. 1 & 2]), wherein the one or more processors execute a program stored in a memory (database 220 can be implemented by any memory, storage device and/or storage disc [¶0065; Fig. 2]) and thereby perform a method comprising: obtaining image data of one frame including a plurality of subjects moving in a specific direction (an image frame depicts a plurality of athletes running along a running track 1152 [¶0148; Fig. 11B]); inputting the image data to a trained machine-learning model; (image data is input into the performance metric determiner 160 comprising trained neural networks [¶0-41-42 & 122; step 412 of Fig. 4]). Leung generally describes tracking a subject of a particular ranking such as a runner in the lead, but fail to describe a particular set of evaluation outputs by which this ranking is determined. Aoyama, however, is analogous art pertinent to the field of endeavor of the present application and disclose an imaging apparatus configured to rank and track subjects in motion. Aoyama teach based on evaluation values output by (Aoyama: based on a face size (i.e., an evaluation value), the subject tracking control unit 153 is able to determine which subject is in the lead in step S211 as they move in a direction approaching the frame of view [¶0053-54; Figs. 2 & 3]); and based on a result of the determination, executing processing for tracking a subject of a specific ranking (Aoyama: based on the ranking of subjects A-E, despite E being closest to the field of view, subject D is tracked as the main subject of tracking as they are detected to be moving in the specific moving direction [¶0052-54; Fig. 3]), and wherein for each of areas of the plurality of subjects, (due to the use of the disjunctive or, only one of the recited limitations requires mapping) as the ranking is closer to the specific ranking (Aoyama: for all subjects A-E in the image, as the face size of a subject increases, the closer they are to the particular ranking of first place [¶0054]). Aoyama explain that it is difficult for photographers to track subjects in scenes with vast and vigorous movements like sports racing, particularly when the ranking of who is in the lead can continuously change overtime [¶0004-5]. The teachings from Aoyama enable real-time subject ranking and tracking based on evaluation metrics like face sizes, distances, or depth information from a point-of-reference. One of ordinary skill in the art before the effective filing date of the present application would recognize that incorporating the teachings of Aoyama with the neural network architecture disclosed by Leung would better enable real-time subject tracking, to ultimately arrive at the invention of the present application. Turning to claim 3, Leung in view of Aoyama teach The image processing apparatus according to claim 1 (as described above), wherein the trained machine-learning model is a trained machine-learning model that uses a neural network, and has been trained to modify parameters of the neural network so that an output of an evaluation function satisfies a convergence condition (Leung: the ML/AI system utilizes a deep neural network model and is trained to iteratively select parameters that reduce a modal error (i.e., a convergence condition) [¶0049-52]), the evaluation function using (i) an output of the neural network corresponding to input image data included in a data set for training, and (ii) supervisory data corresponding to the input image data, as arguments (Leung: supervised training is conducted using training images with corresponding labels (i.e., (ii) supervisory data corresponding to the input image), to output a corresponding expected labelled output (i.e., (i) an output of the neural network) [¶0050-53]). As for claim 8, Leung in view of Aoyama teach The image processing apparatus according to claim 1 (as previously taught), wherein training data set used to train the trained machine-learning model is composed of a pair of input image data and corresponding supervisory data (Leung: training data 235 is comprised of image data provided with corresponding labels [¶0053]), and the supervisory data is in a form of distances to a reference position (Leung: training data 235 can include subjects captured in the image labeled with respective coordinate positions indicating the positions of the subject relative to a captured device view [¶0060]), or values based on the distances (because of the use of the disjunctive “or” only one of the listed limitations requires mapping), as the evaluation values. With respect to claim 9, Leung in view of Aoyama teach The image processing apparatus according to claim 8 (as detailed above), wherein the input image data is generated through 3D modeling of a captured scene (Leung: a 3D keypoint generator 340 uses 2D keypoints extracted from the image along with all related views of a subject of interest to model a scene [¶0109]). Considering claim 12, Leung in view of Aoyama teach The image processing apparatus according to claim 1 (as previously taught), wherein the specific ranking is a first place (Leung: a lead runner (i.e., in first place) is tracked [¶0040 & 60]). Concerning claim 13, Leung teach An image capturing apparatus (Leung: image capture system 110 [¶0037; Fig. 1]), comprising: one or more processors (Leung: the camera system module 140 can be implemented by one or more programmable processors [¶0065; Figs. 1 & 2]), wherein the one or more processors execute a program stored in a memory (Leung: database 220 can be implemented by any memory, storage device and/or storage disc [¶0065; Fig. 2]) and thereby perform a method comprising: generating image data of one frame with use of an image sensor (Leung: the subject detector 310 identifies subjects from the image capture devices by extracting 2D keypoints on a frame-by-frame basis [¶0082-86]); inputting the image data to a trained machine-learning model (Leung: image data is input into the performance metric determiner 160 comprising trained neural networks [¶0-41-42 & 122; step 412 of Fig. 4]); and setting a focus detection area at the detected subject of the specific ranking (Leung: a 3D bounding box 1210 is generated for a first athlete of interest, such as a lead runner (i.e., a specific ranking) [¶0149; Fig. 12A]). Leung generally describes tracking a subject of a particular ranking such as a runner in the lead, but fail to describe a particular set of evaluation outputs by which this ranking is determined. Aoyama, however, teach based on evaluation values output by the trained machine-learning model, detecting a subject of a specific ranking from among a plurality of subjects included in the image data; (Aoyama: based on a face size (i.e., an evaluation value), the subject tracking control unit 153 is able to determine which subject is in the lead in step S211 as they move in a direction approaching the frame of view [¶0053-54; Figs. 2 & 3]). Aoyama explain that it is difficult for photographers to track subjects in scenes with vast and vigorous movements like sports racing, particularly when the ranking of who is in the lead can continuously change overtime [¶0004-5]. The teachings from Aoyama enable real-time subject ranking and tracking based on evaluation metrics like face sizes, distances, or depth information from a point-of-reference. One of ordinary skill in the art before the effective filing date of the present application would recognize that incorporating the teachings of Aoyama with the neural network architecture disclosed by Leung would better enable real-time subject tracking, to ultimately arrive at the invention of the present application. Regarding claim 14, Leung teach An image processing method (Leung: the methods outlined in Figs. 4-10 outlining subject identification, key point extraction, and performance metric determination [¶0118]) executed by an image processing apparatus (Leung: image capture system 110 [¶0037; Fig. 1]), the image processing method comprising: obtaining image data of one frame including a plurality of subjects moving in a specific direction (Leung: an image frame depicts a plurality of athletes running along a running track 1152 [¶0148; Fig. 11B]); inputting the image data to a trained machine-learning model; (Leung: image data is input into the performance metric determiner 160 comprising trained neural networks [¶0-41-42 & 122; step 412 of Fig. 4]). Leung generally describes tracking a subject of a particular ranking such as a runner in the lead, but fail to describe a particular set of evaluation outputs by which this ranking is determined. Aoyama, however, teach based on evaluation values output by (Aoyama: based on a face size (i.e., an evaluation value), the subject tracking control unit 153 is able to determine which subject is in the lead in step S211 as they move in a direction approaching the frame of view [¶0053-54; Figs. 2 & 3]); and based on a result of the determination, executing processing for tracking a subject of a specific ranking (Aoyama: based on the ranking of subjects A-E, despite E being closest to the field of view, subject D is tracked as the main subject of tracking as they are detected to be moving in the specific moving direction [¶0052-54; Fig. 3]), and wherein for each of areas of the plurality of subjects, (due to the use of the disjunctive or, only one of the recited limitations requires mapping) as the ranking is closer to the specific ranking (Aoyama: for all subjects A-E in the image, as the face size of a subject increases, the closer they are to the particular ranking of first place [¶0054]). Aoyama explain that it is difficult for photographers to track subjects in scenes with vast and vigorous movements like sports racing, particularly when the ranking of who is in the lead can continuously change overtime [¶0004-5]. The teachings from Aoyama enable real-time subject ranking and tracking based on evaluation metrics like face sizes, distances, or depth information from a point-of-reference. One of ordinary skill in the art before the effective filing date of the present application would recognize that incorporating the teachings of Aoyama with the neural network architecture disclosed by Leung would better enable real-time subject tracking, to ultimately arrive at the invention of the present application. With respect to claim 15, Leung teach A non-transitory computer-readable medium (Leung: non-transitory computer readable storage medium such as a hard-drive, or memory associated with processors 2012 & 2112 [¶0067]) storing a program which, when executed by a computer (Leung: computing system 225 [¶0067; Fig. 2]), causes the computer to perform an image processing method (Leung: the methods outlined in Figs. 4-10 outlining subject identification, key point extraction, and performance metric determination [¶0118])) comprising: obtaining image data of one frame including a plurality of subjects moving in a specific direction (Leung: an image frame depicts a plurality of athletes running along a running track 1152 [¶0148; Fig. 11B]); inputting the image data to a trained machine-learning model; (Leung: image data is input into the performance metric determiner 160 comprising trained neural networks [¶0-41-42 & 122; step 412 of Fig. 4]). Leung generally describes tracking a subject of a particular ranking such as a runner in the lead, but fail to describe a particular set of evaluation outputs by which this ranking is determined. Aoyama, however, teach based on evaluation values output by (Aoyama: based on a face size (i.e., an evaluation value), the subject tracking control unit 153 is able to determine which subject is in the lead in step S211 as they move in a direction approaching the frame of view [¶0053-54; Figs. 2 & 3]); and based on a result of the determination, executing processing for tracking a subject of a specific ranking (Aoyama: based on the ranking of subjects A-E, despite E being closest to the field of view, subject D is tracked as the main subject of tracking as they are detected to be moving in the specific moving direction [¶0052-54; Fig. 3]), and wherein for each of areas of the plurality of subjects, (due to the use of the disjunctive or, only one of the recited limitations requires mapping) as the ranking is closer to the specific ranking (Aoyama: for all subjects A-E in the image, as the face size of a subject increases, the closer they are to the particular ranking of first place [¶0054]). Aoyama explain that it is difficult for photographers to track subjects in scenes with vast and vigorous movements like sports racing, particularly when the ranking of who is in the lead can continuously change overtime [¶0004-5]. The teachings from Aoyama enable real-time subject ranking and tracking based on evaluation metrics like face sizes, distances, or depth information from a point-of-reference. One of ordinary skill in the art before the effective filing date of the present application would recognize that incorporating the teachings of Aoyama with the neural network architecture disclosed by Leung would better enable real-time subject tracking, to ultimately arrive at the invention of the present application. Claim(s) 2, 4-6 is/are rejected under 35 U.S.C. 103 as being unpatentable over Leung et al. (US 2020/0401793 A1), hereinafter referred to as “Leung”, further in view of Aoyama et al. (US 2022/0005205 A1) hereinafter referred to as “Aoyama”, further in view of Shamir; Gil (WO 2024/063765 A1), hereinafter referred to as “Shamir”. Regarding claim 2, Leung in view of Aoyama teach The image processing apparatus according to claim 1, (taught previously). Leung, while generally describing the use of gradient descent to enable accurate classification from input images, they fail to recite that their machine-learning model’s accuracy increases as a ranking approaches a particular rank. Shamir, on the other hand, describe a method directed to improving ranking using pairwise and listwise ranking losses. Shamir teach wherein the trained machine-learning model has been trained so that accuracy becomes higher as the ranking is closer to the specific ranking (Shamir: individual ordinal CDF (cumulative distribution functions)-point losses are weighted differently so the more critical ranking (i.e., a specific ranking) receive greater weight during training, emphasizing model accuracy for that particular ranking [¶32-36, 45-50] – the examiner notes that this weighting for improved accuracy described by Shamir is in accordance with how accuracy is described to increase with a larger weight used for a higher ranking in ¶0082-86 of the specification). Shamir further explain that certain ordinal rankings are more critical than others and assigns greater loss weighting to those rankings in order to improve overall model ranking performance. It would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to utilize the CDF-point loss weighting of Shamir with the base ranking and tracking architecture taught by Leung in view of Aoyama to arrive at the invention of the invention as claimed in claim 2. This would enable the neural network to more heavily weight rankings as they approach a particular specific or critical ranking. Turning to claim 4, Leung in view of Aoyama teach The image processing apparatus according to claim 3 (as previously taught), but does not describe a weighted addition of ranking evaluation functions. Shamir, per contra, teach wherein the evaluation function is expressed as weighted addition of ranking evaluation functions, and a weight for the ranking evaluation function corresponding to the specific ranking is larger than weights for the ranking evaluation functions corresponding to other rankings (Shamir: a ranking loss function uses separate ordinal CDF-point losses and combines their contributions through a weighted loss function, with greater differential weighting for rankings closer to a critical ranking [¶45-50; eqs. 6 & 7]). Shamir further explain that certain ordinal rankings are more critical than others and assigns greater loss weighting to those rankings in order to improve overall model ranking performance. It would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to utilize the CDF-point loss weighting of Shamir with the base ranking and tracking architecture taught by Leung in view of Aoyama to arrive at the invention of the invention as claimed in claim 2. This would enable the neural network to more heavily weight rankings as they approach a particular specific or critical ranking. As for claim 5, Leung in view of Aoyama teach The image processing apparatus according to claim 3 (as previously taught), but fails to describe ranking evaluation functions for each ranking. Shamir, however, teach wherein the ranking evaluation functions are evaluation functions of the respective rankings (Shamir: separate ranking-loss calculations are applied to each ordinal CDF point, with each loss evaluating a corresponding ranking between the two inputs [¶39-46, 48]. Shamir further explain that certain ordinal rankings are more critical than others and assigns greater loss weighting to those rankings in order to improve overall model ranking performance. It would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to utilize the CDF-point loss weighting of Shamir with the base ranking and tracking architecture taught by Leung in view of Aoyama to arrive at the invention of the invention as claimed in claim 2. This would enable the neural network to more heavily weight rankings as they approach a particular specific or critical ranking. Concerning claim 6, Leung in view of Aoyama teach The image processing apparatus according to claim 3 (previously taught), but fails to describe ranking evaluation functions for each ranking. Shamir, however, teach wherein the ranking evaluation functions are evaluation functions of respective combinations of the rankings (Shamir: a combined score vector 212 is determined for the first (204) and second (208) input based on first (206) and second (210) score vectors [¶42; Fig. 2], with pairwise ranking losses to evaluation relationships between pairs of inputs and listwise ranking losses evaluating relationships between multiple inputs [¶53-56; eqs. 9 & 10]). Shamir further explain that certain ordinal rankings are more critical than others and assigns greater loss weighting to those rankings in order to improve overall model ranking performance. It would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to utilize the CDF-point loss weighting of Shamir with the base ranking and tracking architecture taught by Leung in view of Aoyama to arrive at the invention of the invention as claimed in claim 2. This would enable the neural network to more heavily weight rankings as they approach a particular specific or critical ranking. Claim(s) 7 & 10 is/are rejected under 35 U.S.C. 103 as being unpatentable over Leung et al. (US 2020/0401793 A1), hereinafter referred to as “Leung”, further in view of Aoyama et al. (US 2022/0005205 A1) hereinafter referred to as “Aoyama”, further in view of Pavlakos et al. (“Ordinal Depth Supervision for 3D Human Pose Estimation”, IEEE, 2018), hereinafter referred to as “Pavlakos”. Regarding claim 7, Leung in view of Aoyama teach The image processing apparatus according to claim 1 (as previously detailed), wherein training data set used to train of the machine-learning model is composed of a pair of input image data and corresponding supervisory data, (Leung: Leung: training data 235 is comprised of image data provided with corresponding labels [¶0053]). Leung in view of Aoyama fail to teach utilizing heat maps as a form of supervisory data, instead utilizing labels. Pavlakos, on the other hand, is analogous art pertinent to the field of endeavor of the present application and teach a framework for 3D pose estimation utilizing heatmaps indicating keypoint positions as ground truth. Pavlakos teach and the supervisory data is in a form of a heat map indicating a distribution of the evaluation values (Pavlakos: ground truth (i.e., supervisory data) in the form of heatmaps are used to calculate a loss for predicted keypoint positions [Sec 3.3. Volumetric prediction for 3D pose - ¶05; Fig. 2]). Pavlakos detail that the ground truth heatmaps inform an an L2 loss between predicted keypoints, which enforces overall 3D consistency during supervised training [Sec 3.3. Volumetric prediction for 3D pose - ¶05-06]. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to incorporate the teachings of Pavlakos with the base invention of Leung in view of Aoyama to arrive at the invention of the instant application. Doing so would enable better keypoint estimation and subsequent ranking of subjects in a 3D space. Concerning claim 10, Leung in view of Aoyama, further in view of Pavlakos teach The image processing apparatus according to claim 7 (detailed above), wherein the input image data includes input image data generated by processing another input image data (Leung: an image resolution augmenter 305 enhances image resolution (i.e., processing) to improve image quality for subject and feature detection [¶0123-124; steps 415 & 420 of Fig. 4]), and supervisory data corresponding to the input image data generated by processing the other input image data is generated based on supervisory data corresponding to the other input image data (Leung: training data 362 can include images at different resolutions (such as before and after augmentation) when training the neural network on the labeling task [¶0132-133; Fig. 5B]). Claim(s) 11 is/are rejected under 35 U.S.C. 103 as being unpatentable over Leung et al. (US 2020/0401793 A1), hereinafter referred to as “Leung”, further in view of Aoyama et al. (US 2022/0005205 A1) hereinafter referred to as “Aoyama”, further in view of Pavlakos et al. (“Ordinal Depth Supervision for 3D Human Pose Estimation”, IEEE, 2018), hereinafter referred to as “Pavlakos”, further in view of Cai et al. (“Semi-supervised Active Learning Based on Semantic-aware Crop Consistency”, 2022, IEEE), hereinafter referred to as “Cai”. Considering claim 11, Leung, in view of Aoyama, further in view of Pavlakos teach The image processing apparatus according to claim 10 (detailed above), wherein in a case where at least a part of an area of a subject that has been included in the other input image data is no longer included therein as a result of the processing, (Leung: the 2D keypoint extractor 335 is able to account for scenarios in which some occlusions may occur, such as when an elbow joint of a subject is blocked [¶0108]). Leung while disclosing accounting for keypoints during occluded scenes, they fail to recite utilizing occluded or obscured training images as supervisory data. Cai, however, is analogous art pertinent to the field of endeavor of the present application and disclose an semi-supervised learning strategy for semantic-aware image cropping for training an image classification network. Cai teach the supervisory data corresponding to the input image data generated by processing the other input image data is generated by modifying the supervisory data corresponding to the other input image data in response to the processing (Cai: during image classification training, images are cropped according to a semantic-aware cropping strategy [Sec 1. Introduction – pg. 656; Fig. 1]). Cai explain that image augmentation is used to improve model classification consistency [Sec IIB. Semi-supervised Active Learning - ¶02], with image cropping providing the best overall consistency regularization [Sec C. Data Augmentation - ¶01-02; Table 1]. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to incorporate the teaching of Cai using cropped images to improve the overall scene ranking (a form of image classification) neural network framework taught by Leung in view of Aoyama, further in view of Pavlakos to arrive at the invention of the instant application. Conclusion The prior art made of record and not relied upon is considered pertinent to applicant's disclosure: Zhang et al. (“Monocular Object Instance Segmentation and Depth Ordering with CNNs”, 2015, arXiv) teach an object detection algorithm for estimating the depth of objects from a particular frame of view Freire-Obregon et al. (“A Large-Scale Re-identification Analysis in Sporting Scenarios: the Betrayal of Reaching a Critical Point”, 2023, IEEE) outline a participant identifier using computer vision. Chen et al. (“Ranking Measures and Loss Functions in Learning to Rank”, 2009, NeurIPS) detail the loss functions and ranking measures used during ranking function training. Blondeau et al. (US 2022/0001294 A1) detail a neural network framework for detecting the ranking of subjects as they pass a crossing line. Howell; Nick (US 2017/0236029 A1) describe a method for training a machine learning model to track particular subjects and their associated temporal information as they progress through a race. Any inquiry concerning this communication or earlier communications from the examiner should be directed to Michael M. Sofroniou whose telephone number is (571)272-0287. The examiner can normally be reached M-F: 8:30 AM - 5:00 PM. 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, John M. Villecco can be reached at (571) 272-7319. 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. /MICHAEL M SOFRONIOU/Examiner, Art Unit 2661 /JOHN VILLECCO/Supervisory Patent Examiner, Art Unit 2661
Read full office action

Prosecution Timeline

Feb 12, 2025
Application Filed
Nov 05, 2026
Non-Final Rejection mailed — §101, §103, §112 (current)

Precedent Cases

Applications granted by this same examiner with similar technology

Patent 12784246
INFORMATION PROCESSING APPARATUS, GENERATION METHOD, AND COMPUTER PROGRAM PRODUCT
2y 6m to grant Granted Nov 03, 2026
Patent 12711652
IMAGE PROCESSING APPARATUS, IMAGE PICKUP APPARATUS, CONTROL METHOD FOR IMAGE PROCESSING APPARATUS, AND STORAGE MEDIUM CAPABLE OF NOTIFYING USER OF BLUR INFORMATION, OR OF DISPLAYING INDICATOR
2y 8m to grant Granted Aug 18, 2026
Study what changed to get past this examiner. Based on 2 most recent grants.

Strategy Recommendation AI-generated — please review before filing

Get a prosecution strategy drawn from examiner precedents, rejection analysis, and claim mapping.
Typically takes 5-10 seconds — AI-generated, attorney review required before filing

Prosecution Projections

1-2
Expected OA Rounds
100%
Grant Probability
99%
With Interview (+0.0%)
2y 4m (~7m remaining)
Median Time to Grant
Low
PTA Risk
Based on 5 resolved cases by this examiner. Grant probability derived from career allowance rate.

Sign in with your work email

Enter your email to receive a magic link. No password needed.

Personal email addresses (Gmail, Yahoo, etc.) are not accepted.

Free tier: 3 strategy analyses per month