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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 01/22/2026 has been entered.
Response to Amendment
The amendments filed 01/22/2026 have been entered.
Claims 1-20 remain pending within the application.
Information Disclosure Statement
The information disclosure statements (IDS) submitted on 01/22/2026 and 07/30/2026 are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
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 7 is 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.
Claim 7 recites the limitation "…the second sensor data instances include a first sensor data instance depicting geography…". It is unclear whether this “a first sensor data instances” in claim 7 refers to the “first sensor data instances” recited in claim 1, or another separate sensor data instance. There is insufficient antecedent basis for this limitation in the claim. The examiner notes that no prior art rejection will be given for claim 7 due to the indefinite nature of the claim.
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.
The claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Claim 1 includes the steps of:
At least one processor comprising: one or more circuits to:
determine, based at least on applying first sensor data instances to one or more first machine learning models (MLMs), output data indicating respective probabilities that the first sensor data instances correspond to a first cell of a spatial cell model of an environment, the first sensor data instances depicting respective views of a second cell of the spatial cell model;
aggregating the respective probabilities into a probability that the first sensor data instances depict geography in the first cell based at least on a level of perceptual similarity between the first sensor data instances and perceptual characteristics associated with the first cell;
based at least on the
applying second sensor data instances to one or more second MLMs based at least on associating the second sensor data instances with the semantic region.
Step 2A Prong 1: Does the claim recite an abstract idea, law of nature, or natural phenomenon?
The broadest reasonable interpretation of the following limitations falls within the mental process groupings of abstract ideas because they cover concepts performed in the human mind, including observation, evaluation, judgment, and opinion. See MPEP 2106.04(a)(2), subsection III. The claim(s) recite(s) in part:
“determine, based at least on applying first sensor data instances to one or more first machine learning models (MLMs), output data indicating respective probabilities that the first sensor data instances correspond to a first cell of a spatial cell model of an environment, the first sensor data instances depicting respective views of a second cell of the spatial cell model”. As drafted and under its broadest reasonable interpretation, this limitation recites an abstract idea of a mental process because determining, based on applying first sensor data to machine learning models, output data indicating respective probabilities that the first sensor data instances correspond to a first cell of a spatial cell model of an environment, with the first sensor data instances depicting respective views of a second cell of the spatial cell model, with the models recited at a high level of generality, encompasses mental evaluations that are practically performed in the human mind, but for the recitation of generic computer components. Even if most humans would use a physical aid, like a pen and paper or a calculator, to make such evaluations, the use of a physical aid would not negate the mental nature of this limitation. See MPEP 2106.04(a)(2), subsection III.B.
“aggregating the respective probabilities into a probability that the first sensor data instances depict geography in the first cell based at least on a level of perceptual similarity between the first sensor data instances and perceptual characteristics associated with the first cell”. As drafted and under its broadest reasonable interpretation, this limitation recites an abstract idea of a mental process because aggregating probabilities into a probability that the first sensor data instances depict geography in the first cell based at least on a level of perceptual similarity encompasses mental evaluations that are practically performed in the human mind, but for the recitation of generic computer components. Even if most humans would use a physical aid, like a pen and paper or a calculator, to make such evaluations, the use of a physical aid would not negate the mental nature of this limitation. See MPEP 2106.04(a)(2), subsection III.B.
“based at least on theAs drafted and under its broadest reasonable interpretation, this limitation recites an abstract idea of a mental process because generating a cluster that groups the first cell and the second cell into a semantic region of the environment based on probabilities, recited at a high level of generality, encompasses mental evaluations that are practically performed in the human mind, but for the recitation of generic computer components. Even if most humans would use a physical aid, like a pen and paper or a calculator, to make such evaluations, the use of a physical aid would not negate the mental nature of this limitation. See MPEP 2106.04(a)(2), subsection III.B.
“applying second sensor data instances to one or more second MLMs based at least on associating the second sensor data instances with the semantic region”. As drafted and under its broadest reasonable interpretation, this limitation recites an abstract idea of a mental process because applying second sensor data instances to one or more second MLMs based at least on associating the second sensor data instances with the semantic region, recited at a high level of generality, encompasses mental evaluations that are practically performed in the human mind, but for the recitation of generic computer components. Even if most humans would use a physical aid, like a pen and paper or a calculator, to make such evaluations, the use of a physical aid would not negate the mental nature of this limitation. See MPEP 2106.04(a)(2), subsection III.B.
Step 2A Prong 2: Does the claim recite additional elements that integrate the judicial exception into a practical application?
The judicial exception is not integrated into a practical application. In particular, The claim(s) recite(s) in part:
“At least one processor comprising: one or more circuits to”. As drafted and under its broadest reasonable interpretation, this limitation recites additional elements which amount to generic computer components recited at a high level of generality, with merely the words “apply it” or an equivalent with the judicial exception, merely including instructions to implement an abstract idea on the additional elements, or merely using the additional elements as a tool to perform an abstract idea, as discussed in MPEP 2106.05(f).
The additional elements have been considered both individually and as an ordered combination in order to determine whether they integrates the exception into a practical application. Therefore, no meaningful claim limits are imposed practicing the abstract idea. Accordingly, at Step 2A, prong two, the additional elements do not integrate the judicial exception into a practical application.
Step 2B: Does the claim recite additional elements that amount to significantly more than the judicial exception?
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed, the claim limitations reciting generic computer elements amounts to no more than mere instructions to apply the exception using a generic computer.
The additional elements have been considered both individually and as an ordered combination in order to determine whether they warrant significantly more consideration. Thus, the claim does not provide an inventive concept.
The claim is ineligible.
Claims 2-8
Claims 2-8 further recite limitations that encompass mental evaluations that are practically performed in the human mind, but for the recitation of generic computer components. Claim 8 reciting “a control system for an autonomous or semi-autonomous machine…” merely indicate a field of use or technological environment in which the judicial exception is performed. This type of limitation merely confines the use of the abstract idea to a particular technological environment (autonomous vehicles) and thus fails to add an inventive concept to the claims. See MPEP 2106.05(h). The claims do not integrate the judicial exception into practical application. The claims do not recite additional elements that are sufficient to amount to significantly more than the judicial exception.
Claims 2-8 are ineligible.
Claims 9-15 and 16-20 are substantially similar to claims 1-8, and are rejected on the same basis as claims 1-8. These claims recite additional elements that amount to generic computer components recited at a high level of generality, with merely the words “apply it” or an equivalent with the judicial exception, merely including instructions to implement an abstract idea on the additional elements, or merely using the additional elements as a tool to perform an abstract idea.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-2, 4, 6, 8-13, and 15-20 are rejected under AIA 35 U.S.C. 102(a)(1) as being anticipated by Hu et al. (“Image-Based Geo-Localization Using Satellite Imagery”), hereafter Hu.
Regarding claim 1, Hu discloses:
At least one processor comprising: one or more circuits to (Hu, Fig. 5 and page 9, left column, paragraph 2, lines 1-2 “We implement our CVM-Nets using Tensorflow and train using the Adam optimizer” teaches a processor comprising one or more circuits):
determine, based at least on applying first sensor data instances to one or more first machine learning models (MLMs), output data indicating respective probabilities that the first sensor data instances correspond to a first cell of a spatial cell model of an environment (Hu, Figs. 1, 2, 3 and page 7, left column, paragraph 2, lines 3-5 “The satellite map is discretized into a database of image patches centered on every P pixels of the map.” and page 4, left column, paragraph 1, lines 12-14 “The measurement probability is computed based on the descriptors extracted from our proposed CVM-Net.” teaches applying first sensor data instances to one or more first machine learning models, i.e. applying a ground image to CVM-Nets, to determine output data indicating respective probabilities that the first sensor data instances correspond to a first cell of a spatial cell model of an environment, i.e. a localized area of the satellite map in Fig. 1),
the first sensor data instances depicting respective views of a second cell of the spatial cell model (Hu, Figs. 1, 2, 3 teaches the first sensor data instances to depict respective views of a second cell of the spatial cell model),
aggregating the respective probabilities into a probability that the first sensor data instances depict geography in the first cell based at least on a level of perceptual similarity between the first sensor data instances and perceptual characteristics associated with the first cell (Examiner notes: for prior art purposes, the examiner interprets the BRI of aggregating respective probabilities to include a probability distribution, as per ¶[0024] of the applicant’s disclosure) (Hu, Algorithm 1, page 4, left column, paragraph 1, lines 5-14 “to compute the measurement probability from the ground view query image and the satellite imagery … Our CVM-Net extracts the global descriptors of ground view and satellite images. The descriptor distance indicates the similarity of the ground and satellite images. The measurement probability is computed based on the descriptors extracted from our proposed CVM-Net” and page 7, left column, final paragraph, lines 19-22 “In the particle filter algorithm, the sensor measurement probability distribution…is obtained from the location probability” teaches aggregating the respective probabilities into a probability for a probability distribution that the first sensor data instances depict geography in the first cell based at least on a level of perceptual similarity between the first sensor data instances and perceptual characteristics associated with the first cell),
based at least on the probability, generate a cluster that groups the first cell and the second cell into a semantic region of the environment (Hu, Algorithm 1, Fig. 3 and page 4, right column, paragraph 2, lines 3-7 “NetVLAD [2] is a trainable deep network version of VLAD [17], which aggregates the residuals of the local feature vectors to their respective cluster centroid to generate a global descriptor. The centroids and distance metrics are trainable parameters in NetVLAD.” Teaches generating clusters, based on the probability distribution for cross view matching, that group the first cell and the second cell into a semantic region of the environment, as shown in Fig. 3),
applying second sensor data instances to one or more second MLMs based at least on associating the second sensor data instances with the semantic region (Hu, Figs. 1, 2, and 3 teaches applying a satellite image as a second sensor data instance to one or more second machine learning models, i.e. NetVLADs, based on associating the second sensor data instances with the matching semantic region).
Regarding claim 2, Hu discloses the processor of claim 1. Hu further discloses:
wherein the one or more first MLMs are trained to classify sensor data inputs with corresponding cell identifiers that represent which corresponding cell of the spatial cell model the sensor data inputs depict (Fig. 1 and Fig. 2 teaches the MLMs trained to classify sensor data inputs with localized cell identifiers that represent which corresponding cell of the spatial cell model the sensor data inputs depict),
the probability is that the one or more first MLMs will classify each of the first sensor data instances with a cell identifier of the corresponding cell identifiers that represents the first cell (Hu, Figs. 10 - 11 and page 12, left column, final paragraph, lines 15-17 “We visualize the heatmap of the similarity scores on the reference satellite map of two examples” teaches the probability is that the first machine learning models will classify each sensor data with a cell identifier of the corresponding cell identifiers that represents the first cell).
Regarding claim 4, Hu discloses the processor of claim 1. Hu further discloses:
wherein the spatial cell model represents a partitioning of a geographic region into a plurality of cells that each represent a distinct region of the geographic region, the first cell and the second cell included in the plurality of cells (Hu, page 7, left column, paragraph 2, lines 1-5 “We perform image-based ground-to-satellite geo-localization with respect to a geo-referenced satellite map with our cross-view image retrieval CVM-Nets. The satellite map is discretized into a database of image patches centered on every P pixels of the map.” Teaches the satellite map to be discretized to indicate a partitioning of a geographic region into a plurality of cells, comprising the first and second cell, that each represent a distinct region of the geographic region).
Regarding claim 6, Hu discloses the processor of claim 1. Hu further discloses:
wherein the first cell corresponds to a first country and the second cell corresponds to a second country different from the first country (Hu, page 12, left column, paragraph 3, lines 7-10 “We choose to test our CVM-Net on Singapore to show that our CVM-Net trained on the North American based CVUSA datasets generalize well on a drastically different area.” Teaches using cells corresponding to different countries).
Regarding claim 8, Hu discloses the processor of claim 1. Hu further discloses:
wherein the processor is comprised in at least one of: a control system for an autonomous or semi-autonomous machine; a perception system for an autonomous or semi-autonomous machine; a system for performing simulation operations; a system for performing digital twin operations; a system for performing light transport simulation; a system for performing collaborative content creation for 3D assets; a system for performing deep learning operations; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational Al operations; a system for generating synthetic data; a system incorporating one or more virtual machines (VMs); a system implemented at least partially in a data center; or a system implemented at least partially using cloud computing resources (Hu, Fig. 5 teaches a control system for an autonomous or semi-autonomous machine).
Regarding claim 9, Hu discloses:
A system comprising: one or more sensors; and one or more processing units comprising processing circuitry to (Hu, Fig. 5, Fig. 1 element “visual odometry”, page 8, left column, paragraph below equation 9, line 3 “…the sensor measurement…” teaches a system comprising one or more sensor, and page 9, left column, paragraph 2, lines 1-2 “We implement our CVM-Nets using Tensorflow and train using the Adam optimizer” teaches and one or more processing units comprising processing circuitry),
determine, based at least on applying first sensor data instances to one or more first machine learning models (MLMs), output data indicating respective probabilities that the first sensor data instances correspond to a first cell of a spatial cell model of an environment (Hu, Figs. 1, 2, 3 and page 7, left column, paragraph 2, lines 3-5 “The satellite map is discretized into a database of image patches centered on every P pixels of the map.” and page 4, left column, paragraph 1, lines 12-14 “The measurement probability is computed based on the descriptors extracted from our proposed CVM-Net.” teaches applying first sensor data instances to one or more first machine learning models, i.e. applying a ground image to CVM-Nets, to determine output data indicating respective probabilities that the first sensor data instances correspond to a first cell of a spatial cell model of an environment, i.e. a localized area of the satellite map in Fig. 1)
the first sensor data instances depicting respective views of a second cell of the spatial cell model (Hu, Figs. 1, 2, 3 teaches the first sensor data instances to depict respective views of a second cell of the spatial cell model),
aggregate the respective probabilities into a probability that the first sensor data instances depict geography in the first cell (Examiner notes: for prior art purposes, the examiner interprets the BRI of aggregating respective probabilities to include a probability distribution, as per ¶[0024] of the applicant’s disclosure) (Hu, Algorithm 1, page 4, left column, paragraph 1, lines 5-14 “to compute the measurement probability from the ground view query image and the satellite imagery … Our CVM-Net extracts the global descriptors of ground view and satellite images. The descriptor distance indicates the similarity of the ground and satellite images. The measurement probability is computed based on the descriptors extracted from our proposed CVM-Net” and page 7, left column, final paragraph, lines 19-22 “In the particle filter algorithm, the sensor measurement probability distribution…is obtained from the location probability” teaches aggregating the respective probabilities into a probability for a probability distribution that the first sensor data instances depict geography in the first),
generate, based at least on the probability, at least one cluster including a cluster that groups the first cell and the second cell into a semantic region of the environment (Hu, Algorithm 1, Fig. 3 and page 4, right column, paragraph 2, lines 3-7 “NetVLAD [2] is a trainable deep network version of VLAD [17], which aggregates the residuals of the local feature vectors to their respective cluster centroid to generate a global descriptor. The centroids and distance metrics are trainable parameters in NetVLAD.” Teaches generating clusters, based on the probability distribution for cross view matching, that group the first cell and the second cell into a semantic region of the environment, as shown in Fig. 3),
apply second sensor data instances to one or more second MLMs based at least on associating the second sensor data instances with the semantic region (Hu, Figs. 1, 2, and 3 teaches applying a satellite image as a second sensor data instance to one or more second machine learning models, i.e. NetVLADs, based on associating the second sensor data instances with the matching semantic region).
Regarding claim 10, Hu discloses the system of claim 9. Hu further discloses:
output data comprises, for each instance of the first sensor data instances, a probability distribution indicating corresponding probabilities that the instance classified to the first cell by the one or more first machine learning models may also be classified to each other cell of the spatial cell model (Hu, page 2, left column, final paragraph, lines 8-12 “The measurement probability for one ground view image of the Markov localization framework is the probability distribution on the satellite map, which is obtained from the descriptor distances.” Teaches output data to comprise a probability distribution indicating corresponding probabilities that the instance classified to the first cell by the one or more first machine learning models may also be classified to each other cell of the spatial cell model).
Regarding claim 11, Hu discloses the system of claim 9. Hu further discloses:
wherein the spatial cell model indicates a partitioning of a geographic region into a plurality of cells that each represent distinct portions of the geographic region (Hu, page 7, left column, paragraph 2, lines 1-5 “We perform image-based ground-to-satellite geo-localization with respect to a geo-referenced satellite map with our cross-view image retrieval CVM-Nets. The satellite map is discretized into a database of image patches centered on every P pixels of the map.” Teaches the satellite map to be discretized to indicate a partitioning of a geographic region into a plurality of cells that each represent distinct portions of the geographic region).
Regarding claim 12, Hu discloses the system of claim 9. Hu further discloses:
wherein the probability is that the one or more first MLMs will classify the first sensor data instances as depicting geography in the first cell (Hu, Figs. 10 - 11 and page 12, left column, final paragraph, lines 15-17 “We visualize the heatmap of the similarity scores on the reference satellite map of two examples” teaches the probability is that the first machine learning models will classify each sensor as depicting geography in the first cell).
Regarding claim 13, Hu discloses the system of claim 9. Hu further discloses:
wherein the cluster is generated based at least on the probability being within a threshold range (Hu, algorithm 1, algorithm 2, and paragraph below algorithm 1 “The framework first computes the measurement probability (denoted as Satellite Localization) and the state transition probability (denoted as VIO Localization). The new particles are computed through the particle filter algorithm (denoted as PF).” Teaches generating clusters based on probability being within a threshold range through particle filtering).
Regarding claim 15, Hu discloses the system of claim 9. Hu further discloses:
wherein the system is comprised in at least one of: a control system for an autonomous or semi-autonomous machine; a perception system for an autonomous or semi-autonomous machine; a system for performing simulation operations; a system for performing digital twin operations; a system for performing light transport simulation; a system for performing collaborative content creation for 3D assets; a system for performing deep learning operations; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational Al operations; a system for generating synthetic data; a system incorporating one or more virtual machines (VMs); a system implemented at least partially in a data center; or a system implemented at least partially using cloud computing resources (Hu, Fig. 5 teaches a control system for an autonomous or semi-autonomous machine).
Regarding claim 16, Hu discloses:
A method comprising: performing one or more operations by an ego-machine in a first geographic region based at least on an output of a first machine learning model (MLM), the output being generated based at least on (Hu, Fig. 2 and page 12, right column, paragraph 4, lines 1-3 “our image-based cross-view geo-localization framework executed live on the vehicle” and page 12, right column, paragraph 3, lines 1-3 “We perform the real-world experiment in two areas of Singapore - One North and South Buona Vista.” teaches an ego machine, i.e., vehicle, in a geographic region performing operations based on the output of the CVM-Net machine learning model),
determining, based at least on applying first sensor data instances to a MLM, output data indicating respective probabilities that the first sensor data instances correspond to a first cell of a spatial cell model of an environment(Hu, Figs. 1, 2, 3 and page 7, left column, paragraph 2, lines 3-5 “The satellite map is discretized into a database of image patches centered on every P pixels of the map.” and page 4, left column, paragraph 1, lines 12-14 “The measurement probability is computed based on the descriptors extracted from our proposed CVM-Net.” teaches applying first sensor data instances to one or more first machine learning models, i.e. applying a ground image to CVM-Nets, to determine output data indicating respective probabilities that the first sensor data instances correspond to a first cell of a spatial cell model of an environment, i.e. a localized area of the satellite map in Fig. 1)
the first sensor data instances depicting respective views of a second cell of the spatial cell model (Hu, Figs. 1, 2, 3 teaches the first sensor data instances to depict respective views of a second cell of the spatial cell model),
aggregating the respective probabilities into a probability that the first sensor data instances depict geography in the first cell (Examiner notes: for prior art purposes, the examiner interprets the BRI of aggregating respective probabilities to include a probability distribution, as per ¶[0024] of the applicant’s disclosure) (Hu, Algorithm 1, page 4, left column, paragraph 1, lines 5-14 “to compute the measurement probability from the ground view query image and the satellite imagery … Our CVM-Net extracts the global descriptors of ground view and satellite images. The descriptor distance indicates the similarity of the ground and satellite images. The measurement probability is computed based on the descriptors extracted from our proposed CVM-Net” and page 7, left column, final paragraph, lines 19-22 “In the particle filter algorithm, the sensor measurement probability distribution…is obtained from the location probability” teaches aggregating the respective probabilities into a probability for a probability distribution that the first sensor data instances depict geography in the first),
based at least on the probability, generating a cluster that groups the first cell and the second cell into a semantic region of the environment (Hu, Algorithm 1, Fig. 3 and page 4, right column, paragraph 2, lines 3-7 “NetVLAD [2] is a trainable deep network version of VLAD [17], which aggregates the residuals of the local feature vectors to their respective cluster centroid to generate a global descriptor. The centroids and distance metrics are trainable parameters in NetVLAD.” Teaches generating clusters, based on the probability distribution for cross view matching, that group the first cell and the second cell into a semantic region of the environment, as shown in Fig. 3),
applying second sensor data instances to the first MLM based at least on associating the second sensor data instances with the semantic region (Hu, Figs. 1, 2, and 3 teaches applying a satellite image as a second sensor data instance to one or more second machine learning models, i.e. NetVLADs, based on associating the second sensor data instances with the matching semantic region).
Regarding claim 17, Hu discloses the method of claim 16. Hu further discloses:
wherein the spatial cell model indicates a partitioning of one or more geographic regions into a plurality of cells that each are associated with distinct portions of the one or more geographic regions (Hu, page 7, left column, paragraph 2, lines 1-5 “We perform image-based ground-to-satellite geo-localization with respect to a geo-referenced satellite map with our cross-view image retrieval CVM-Nets. The satellite map is discretized into a database of image patches centered on every P pixels of the map.” Teaches the satellite map to be discretized to indicate a partitioning of a geographic region into a plurality of cells that each are associated with distinct portions of the geographic region).
Regarding claim 18, Hu discloses the method of claim 16. Hu further discloses:
wherein the first cell and the second cell are identified as similar based at least in part on a similarity matrix comprising pairwise similarity values corresponding to a plurality of cells, the similarity matrix populated using one or more outputs of the second machine learning model (Hu, page 12, left paragraph last line – right column first 2 lines “our street-view based query image generally return higher similarity scores on areas that correspond to the roads on the satellite map” and page 4, right column, penultimate paragraph, 10-14 “Each NetVLAD layer produces a VLAD vector, i.e. global descriptor, for the respective views vs and vg that are in the same space, which can then be used for direct similarity comparison” teaches a global descriptor used for a similarity matrix comprising a pairwise similarity corresponding to the first cell and the second cell).
Regarding claim 19, Hu discloses the method of claim 16.
wherein the one or more operations are performed based at least on associating the first geographic region with the first cell and based at least on the first cell being included in the semantic region (Hu, Figs. 1-3 and Fig. 5 teaches performing operations based on associations between geographic region and the first cell based on the first cell being included in the semantic region).
Regarding claim 20, Hu discloses the method of claim 16. Hu further discloses:
wherein the method is executed using at least one of: a control system for an autonomous or semi-autonomous machine; a perception system for an autonomous or semi-autonomous machine; a system for performing simulation operations; a system for performing digital twin operations; a system for performing light transport simulation; a system for performing collaborative content creation for 3D assets; a system for performing deep learning operations; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational Al operations; a system for generating synthetic data; a system incorporating one or more virtual machines (VMs); a system implemented at least partially in a data center; or a system implemented at least partially using cloud computing resources (Hu, Fig. 5 teaches a control system for an autonomous or semi-autonomous machine).
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.
Claims 3 is rejected under 35 U.S.C. 103 as being unpatentable over Hu et al. (“Image-Based Geo-Localization Using Satellite Imagery”), hereafter Hu, in view of Chidlovskii et al. (US 2014/0089036 A1), hereafter Xerox.
Regarding claim 3, Hu discloses the processor of claim 1. Hu further discloses:
wherein the probability is included in a similarity … corresponding to a respective pairwise probability that sensor data depicting one or more first cells of the spatial cell model will be classified as corresponding to one or more second cells of the spatial cell model according to the one or more first machine learning models (Hu, page 2, left column, last paragraph, lines 6-12 “the descriptor distance of a ground view image to all the positions on the satellite map can be computed…probability distribution on the satellite map, which is obtained from the descriptor distances.” and page 12, left paragraph last line – right column first 2 lines “our street-view based query image generally return higher similarity scores on areas that correspond to the roads on the satellite map” and page 4, right column, penultimate paragraph, 10-14 “Each NetVLAD layer produces a VLAD vector, i.e. global descriptor, for the respective views vs and vg that are in the same space, which can then be used for direct similarity comparison” teaches through the descriptor distance of ground view images to positions on the satellite map, that the probability is included in a similarity corresponding to a respective pairwise probability that sensor data depicting one or more first cells of the spatial cell model will be classified as corresponding to one or more second cells of the spatial cell model according to the one or more machine learning models).
While Hu teaches wherein the probability is included in a similarity … corresponding to a respective pairwise probability that sensor data depicting one or more first cells of the spatial cell model will be classified as corresponding to one or more second cells of the spatial cell model according to the one or more first machine learning models, they do not explicitly teach the similarity to be a similarity matrix where each cell of the similarity matrix corresponding to a respective pairwise.
Xerox teaches:
a similarity matrix, each cell of the similarity matrix corresponding to a respective pairwise …(Xerox, ¶[0097], ¶[0047], and Fig. 4 element S110 teaches computing a similarity, i.e. affinity, matrix where each cell of the similarity matrix corresponding to a respective pairwise value).
Hu and Xerox are analogous art because they are from the same field of endeavor, image processing and machine learning models.
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Hu to include a similarity matrix, each cell of the similarity matrix corresponding to a respective pairwise, based on the teachings of Xerox. One of ordinary skill in the art would have been motivated to make this modification in order for a good performance on … clusters .. as well as implementation ease (Xerox, ¶[0099]).
Claims 5 is rejected under 35 U.S.C. 103 as being unpatentable over Hu et al. (“Image-Based Geo-Localization Using Satellite Imagery”), hereafter Hu, in view of Vo et al. ("Revisiting IM2GPS in the Deep Learning Era") as recited in the IDS dated 09/02/2022, hereafter Vo.
Regarding claim 5, Hu discloses the processor of claim 4.
Hu teaches partitioning of geographic regions, but does not explicitly teach the partitioning to be determined based on a road network density associated with the geographic region.
Vo discloses:
wherein the partitioning is determined based on a road network density associated with the geographic region (Vo, page 2, right column, fourth point “Through extensive experimentation, we find that some training procedures lead to higher accuracy at the street scale (1km).” and page 3, right column, penultimate paragraph, lines 1-2 “divides the Earth into a set of geographical cells based on image density” teaches partitioning to be based on street, i.e., road network density).
Hu and Vo are analogous art because they are from the same field of endeavor, image processing and machine learning models.
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Hu to include partitioning is determined based on a road network density associated with the geographic region, based on the teachings of Vo. One of ordinary skill in the art would have been motivated to make this modification in order to “improve the state-of-the-art accuracy for global image geolocalization.” (Vo, page 2, right column, first point, lines 1-2).
Claims 14 is rejected under 35 U.S.C. 103 as being unpatentable over Hu et al. (“Image-Based Geo-Localization Using Satellite Imagery”), hereafter Hu, in view of Noda et al. ("Vehicle Ego-Localization by Matching In-Vehicle Camera Images to an Aerial Image"), hereafter Noda.
Regarding claim 14, Hu discloses the system of claim 9. Hu further discloses:
wherein the second sensor data instances depict one or more road … of a first type corresponding to the first cell and, based at least on the cluster, the one or more road … are used to train the one or more second machine learning models to detect road … of a second type that corresponds to the second cell (Hu, Figs. 1-3 and page 2, right column, final paragraph, lines 15-16 “the geo-localization results of a vehicle moving on the road.” teaches the second sensor data instances depict one or more roads of a first type corresponding to the first cell and, based at least on the cluster, the one or more roads are used to train the one or more second machine learning models to detect roads of a second type that corresponds to the second cell).
While Hu teaches wherein the second sensor data instances depict one or more road … of a first type corresponding to the first cell and, based at least on the cluster, the one or more road … are used to train the one or more second machine learning models to detect road … of a second type that corresponds to the second cell, they do not explicitly teach these roads to be road signs.
Noda teaches:
depict one or more road signs of a first type … the one or more road signs are used to train … machine learning models to detect road signs (Noda, Fig. 6, Fig. 4 and page 164, paragraph 2, lines 9-11 “As feature-points, the centroids of road markings, which are traffic symbols printed on roads, are used.” And page 165, first 2 lines “we extract feature-points appropriate for the matching from each road-marking region” teaches depicting one or more road signs of a first type where the one or more road signs are used to train machine learning models to detect road signs for matching).
Hu and Noda are analogous art because they are from the same field of endeavor, image processing and machine learning models.
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the teachings of Hu to include depict one or more road signs of a first type … the one or more road signs are used to train … machine learning models to detect road signs, based on the teachings of Noda. One of ordinary skill in the art would have been motivated to make this modification in order to improve the accuracy and stability (Noda, page 168, paragraph 2, line 4).
Response to Arguments
Applicant's arguments filed 01/22/2026 have been fully considered with regards to the 35 U.S.C. 101 rejection, but they are not persuasive.
The applicant asserts on page 12 of the remarks that the claimed invention fully comply with the requirements of 35 USC 101 and disclose enhancing sensor data and improving performance. The MPEP 2106.04(d)(1) discloses the evaluation of claimed improvements in the functioning of a computer or improvement to a technical field in step 2A prong two. The MPEP section discloses — "if the specification explicitly sets forth an improvement but in a conclusory manner (i.e., a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology. Second, if the specification sets forth an improvement in technology, the claim must be evaluated to ensure that the claim itself reflects the disclosed improvement. That is, the claim includes the components or steps of the invention that provide the improvement described in the specification ... ". The Examiner notes that the claims do not include components or steps of the invention that provide an improvement, as the amended claims are directed to abstract ideas without significantly more or being integrated into a practical application. See 101 rejection above for details.
Applicant's arguments filed 01/22/2026 have been fully considered with regards to the 35 U.S.C. 102/103 rejection, but they are not persuasive.
The applicant asserts on page 13 of the remarks “Applicant respectfully submits that Hu does not disclose, at least, "aggregating the respective probabilities into a probability that the first sensor data instances depict geography in the first cell based at least on a level of perceptual similarity between the first sensor data instances and perceptual characteristics associated with the first cell," and "based at least on the probability, generate a cluster that groups the first cell and the second cell into a semantic region of the environment," as recited in amended claim 1.”
As a preliminary matter, the examiner notes that the disclosure appears to be silent on aggregating respective probabilities beyond the recitation of a probability distribution, as per ¶[0024] of the applicant’s disclosure. Thus, for prior art purposes, the examiner has interpreted the BRI of aggregating respective probabilities to include a probability distribution, which is taught by Hu (Algorithm 1, page 7, left column, final paragraph, lines 19-22 “In the particle filter algorithm, the sensor measurement probability distribution…is obtained from the location probability” and page 4, left column, paragraph 1, lines 5-14 “to compute the measurement probability from the ground view query image and the satellite imagery … Our CVM-Net extracts the global descriptors of ground view and satellite images. The descriptor distance indicates the similarity of the ground and satellite images. The measurement probability is computed based on the descriptors extracted from our proposed CVM-Net”), where a probability distribution teaches the aggregated probability that the first sensor data instances depict geography in the first cell based at least on a level of perceptual similarity between the first sensor data instances and perceptual characteristics associated with the first cell. See 102 rejection above for further details.
Claims 9 and 16, being substantially similar to claim 1, are similarly rejected.
Claims dependent on independent claims do not overcome the deficiencies of the rejected independent claims.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's
disclosure.
- U.S. Pub No. 20230192121 A1: Zhang et al. teaches semantic segmentation and clustering.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HUMAIRA ZAHIN MAUNI whose telephone number is (703)756-5654. The examiner can normally be reached Monday - Friday, 9 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, MATT ELL can be reached at (571) 270-3264. 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.
/H.Z.M./Examiner, Art Unit 2141
/MATTHEW ELL/Supervisory Patent Examiner, Art Unit 2141
