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 .
Drawings
The drawings were received on 10/25/2024. These drawings are accepted.
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)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1-5, 8-15 and 18-20 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Ravi Kumar et al. (US20250078294, hereinafter “Ravi Kumar”)
Claim 1. Ravi Kumar teaches An autonomy computing system of an autonomous vehicle ([0019] “self-drive autonomous vehicle 102”) for feature alignment ([0077] “first feature map…second feature map…outputs of the two attention operations, A1 and A2, may be concatenated”) in multimodal ([0039] “input data 210 may include multimodal data.”) fusion, ([0077] “bi-fusion operation 406 between two feature maps”) comprising at least one processor in communication with at least one memory device, ([0032] “computing system 200 (e.g., processing circuitry 243, memory 202, task networks 206, segmentation decoder 252, etc.) may be interconnected to enable inter-component communications (physically, communicatively, and/or operatively).”) and the at least one processor programmed to: ([0032] “Processing circuitry 243 of computing system 200 may implement functionality and/or execute instructions associated with computing system 200.”)
receive a first feature map ([0077] “first feature map 402”) extracted from first sensor data of an environment ([0051] “(images acquired by at least one of the plurality of cameras 130-134) to determine the distance of objects in the environment.”) and a second feature map ([0077] “second feature map 404.”) extracted from second sensor data of the environment, ([0051] “(images acquired by at least one of the plurality of cameras 130-134) to determine the distance of objects in the environment.”) wherein the autonomous vehicle is operating in the environment, ([0016] “In some cases, an autonomous vehicle may need to perform automated parking/driving after heavy rain, for example, in a parking lot.”) the first sensor data being from one or more sensors of a first modality ([0022] “array of sensor inputs including, for example: one or more ultrasonic sensors 124,”) and the second sensor data being from one or more sensors of a second modality, ([0022] “one or more Light Detection and Ranging (“LIDAR”) sensors 128,”) the one or more sensors of the first modality and the one or more sensors of the second modality installed on the autonomous vehicle; ([0022] “such cameras are located at various places on vehicle body 104”)
fuse the first feature map and the second feature map into a fused feature map ([0077] “bi-fusion operation 406 between two feature maps”) by:
associating first cells ([0101] “use the uncertainty estimates 424 to identify pixels”) in the first feature map with second cells in the second feature map ([0101] “concatenate 428 the desired feature map 426”) based on at least one of uncertainty in the first feature map or uncertainty in the second feature map; ([0101] “with the uncertainty estimates 424 of the multi-class segmentation feature map 404.”) and
determining the fused feature map based on attention([0077] “outputs of the two attention operations, A1 and A2, may be concatenated”) between the first feature map([0077] “first feature map 402”) and the second feature map([0077] “second feature map 404.”) among associated cells; ([0098] “fused feature map 420 to classify each pixel”) and
control operation of the autonomous vehicle ([0019] “output autonomous operation commands to self-drive autonomous vehicle 102”) based on the fused feature map. ([0106] “feature map 414 may be used to estimate the road surface…The uncertainty estimates 424 may be used to determine…the speed of the autonomous vehicle 102”)
Claim 2. Ravi Kumar teaches The autonomy computing system of claim 1, wherein the at least one processor is further programmed to:
associate the first cells by:
for a query cell among the first cells, ([0077] “may perform the bi-fusion operation 406 between two feature maps F1 and F2 by performing the following steps:” and [0078] “1.Select a query from F1”) associating the query cell with key cells in the second feature map, ([0079] “2. Select keys and values from F2.”) wherein the key cells correspond to the query cell and neighboring cells of the query cell in one or more regions ([0080] “3. Apply attention to the keys and values, using the query as the reference.” And [0081] “4. The output of the attention operation is a weighted sum of the values, where the weights are determined by the attention scores.” And [0084] “attention mechanism may include a function to calculate the attention scores between the query and the keys.”) determined based on at least one of the uncertainty in the first feature map or the uncertainty in the second feature map. ([0101] “concatenate 428 the desired feature map 426 with the uncertainty estimates 424 of the multi-class segmentation feature map 404. The uncertainty estimates 424 may be used by the machine learning system 204 to weight the desired feature map 426.”)
Claim 3. Ravi Kumar teaches The autonomy computing system of claim 1, wherein the at least one processor is further programmed to:
determine the fused feature map by:
computing the attention ([0080] “apply attention”) between the first feature map ([0077] “first feature map 402”) and the second feature map ([0077] “second feature map 404.”) among the associated cells,([0085] “pixel”) wherein queries are based on query cells in one modality ([0078] “1.Select a query from F1”) and keys and values are based on key cells in the other modality ([0079] “2. Select keys and values from F2.” Each feature map is associated with another modality as cited in the above rejections of claims 1 and 2) associated with the query cells. ([0080] “3. Apply attention to the keys and values, using the query as the reference.”)
Claim 4. Ravi Kumar teaches The autonomy computing system of claim 1, wherein the first sensor data is two-dimensional (2D), ([0051] “image from the camera” is known to be the same as the claimed sensor data is two dimensional in light of instant specifications [0037] ) the at least one processor further programmed to:
compute depth information of the first sensor data; ([0051] “depth estimation model 206A may be configured to use input data 210 (images acquired by at least one of the plurality of cameras 130-134)”) and
determine depth uncertainty based on the depth information. ([0099] “The uncertainty estimates 424 of the multi-class segmentation feature map 415”)
Claim 5. Ravi Kumar teaches The autonomy computing system of claim 4, wherein the at least one processor is further programmed to:
determine the uncertainty in the first feature map ([0099] “The uncertainty estimates 424 of the multi-class segmentation feature map 415 may provide information about the confidence of the segmentation algorithm.”) based on the depth uncertainty. ([0085] “using the depth information to weight the segmentation predictions.”)
Claim 8. Ravi Kumar teaches The autonomy computing system of claim 1, wherein the at least one processor is further programmed to:
compute the uncertainty in the first feature map ([0101] “feature map 426 with the uncertainty estimates 424”) as a weighted sum of uncertainty determined online ([0103] “fine-tuning” is understood to be the same as the claimed fine-tuning in light of instant specifications [0039] ) and uncertainty based on offline calibration. ([0103] “Fine-tuning is a technique that may be used to improve the performance of a pre-trained model.” Pre-trained model is understood to be the same as the claimed offline calibration in light of instant specifications [0039])
Claim 9. Ravi Kumar teaches The autonomy computing system of claim 1, wherein the at least one processor is further programmed to:
concatenate a context feature map based on the attention ([0083] “The outputs of the two attention operations, A1 and A2, may be concatenated and fed forward.”) and a query feature map into the fused feature map, ([0077] “may perform the bi-fusion operation 406 between two feature maps F1 and F2 by performing the following steps:”) the query feature map being at least one of the first feature map or the second feature map ([0078] “Select a query from F1. [0079] 2. Select keys and values from F2.”) used as queries in computing the attention. ([0080 ] “Apply attention to the keys and values, using the query as the reference.”)
Claim 10. Ravi Kumar teaches The autonomy computing system of claim 1, wherein the first sensor data is two-dimensional (2D), ([0051] “image from the camera” is known to be the same as the claimed sensor data is two dimensional in light of instant specifications [0037] ) the at least one processor further programmed to:
estimate depth information of the first sensor data by:
estimating first depth information using a first mechanism; ([0095] “Machine learning system 204 may create the height feature map 410 by converting the height information in the image to a grayscale image.”)
estimating second depth information using a second mechanism; ([0095] “Machine learning system 204 may create the depth variance feature map 402 by calculating the variance of the depth values in a small neighborhood around each pixel.”) and
fusing the first depth information and the second depth information into the depth information of the first sensor data. ([0095] “perform the first tri-fusion operation 408”)
Claim 11. The method herein has been executed and performed by the system of claim 1 and is likewise rejected
Claim 12. The method herein has been executed and performed by the system of claim 2 and is likewise rejected
Claim 13. The method herein has been executed and performed by the system of claim 3 and is likewise rejected
Claim 14. The method herein has been executed and performed by the system of claim 4 and is likewise rejected
Claim 15. The method herein has been executed and performed by the system of claim 5 and is likewise rejected
Claim 18. The method herein has been executed and performed by the system of claim 8 and is likewise rejected
Claim 19. The method herein has been executed and performed by the system of claim 9 and is likewise rejected
Claim 20. The method herein has been executed and performed by the system of claim 10 and is likewise rejected
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 for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 6-7 and 16-17 are rejected under 35 U.S.C. 103 as being unpatentable over Ravi Kumar et al. (US20250078294, hereinafter “Ravi Kumar”) and in view of Durasov et al (US20260111771, hereinafter “Durasov”)
Claim 6. Ravi Kumar teaches The autonomy computing system of claim 4, wherein the at least one processor is further programmed to:
Ravi Kumar does not explicitly teach determine the uncertainty in the first feature map by applying an unscented transformation to the depth uncertainty.
Durasov teaches determine the uncertainty in the first feature map ([0031] “The uncertainty estimates may be generated for each class and cell of the representation of features corresponding to a scene.”) by applying an unscented transformation to the depth uncertainty. ([0249] “ The generative LM 930 and/or other components of the generative LM system 900 may use different types of neural network architectures depending on the implementation. For example …applying non-linear transformations to the input representations and extracting higher-level features.” is understood to be the same as the claimed applying an unscented transformation in light of instant specifications [0057] )
It would have been obvious to persons of ordinary skill in the art before the effective filing date of the claimed invention to modify Ravi Kumar to have an unscented transformation to the feature map depth uncertainty as taught by Durasov to arrive at the claimed invention discussed above. The motivation for the proposed modification would have been so that (Durasov et al [0002]“multi-modal approaches enable the model to leverage the complementary strengths of a camera and LiDAR, which yields improved detection accuracy over single modality methods.”)
Claim 7. Ravi Kumar teaches The autonomy computing system of claim 4, wherein the at least one processor is further programmed to:
Ravi Kumar does not explicitly teach determine the depth uncertainty as statistics of a probability distribution of the depth uncertainty as a Gaussian distribution.
Durasov teaches determine the depth uncertainty as statistics of a probability distribution of the depth uncertainty as a Gaussian distribution. ([0029] “uncertainty estimates… representation of features where an actual object center is located, and a Bayes risk loss may be computed for each of these cells and scaled (e.g., using a Gaussian Focal Loss (GFL)-based factor),” is understood to be the same as a gaussian distribution because a Gaussian Focal Loss utilizes a Gaussian distribution as per NPL “An Improved STARK Tracker with Gaussian Distributed Focal Loss.” by Zhu et al )
It would have been obvious to persons of ordinary skill in the art before the effective filing date of the claimed invention to modify Ravi Kumar to have uncertainty as a gaussian distribution as taught by Durasov to arrive at the claimed invention discussed above. The motivation for the proposed modification would have been so that (Durasov et al [0002]“multi-modal approaches enable the model to leverage the complementary strengths of a camera and LiDAR, which yields improved detection accuracy over single modality methods.”)
Claim 16. The method herein has been executed and performed by the system of claim 6 and is likewise rejected
Claim 17. The method herein has been executed and performed by the system of claim 7 and is likewise rejected
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant’s disclosure:
Zhu et al NPL "An Improved STARK Tracker with Gaussian Distributed Focal Loss." Discloses utilizing a gaussian distribution to employ attention mechanism for object tracking by capturing features in a region
Li et al US20250315932 discloses sub-pixel registration to achieve precise alignment between frames of multimodal sensors such as RGB camera, LIDAR, thermal cameras etc.
Plaut et al US20250313228 discloses uncertainty calculations when fusing Camera and Lidar data to classify objects according to a combined confidence score.
Hwang et al US20230213643 discloses generating a feature representation of the fused camera and radar/LIDAR data using a neural network with a local attention mechanism.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to OWAIS MEMON whose telephone number is (571)272-2168. The examiner can normally be reached M-F (7:00am - 4:00pm) CST.
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, Gregory Morse can be reached at (571) 272-3838. 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.
/OWAIS I MEMON/Examiner, Art Unit 2663
Prosecution Insights