IP Library Granted Patent US 12,198,059
Granted Patent B2
US 12,198,059 · App. 18/207,788 · Granted Jan 14, 2025

Systems and methods for performing direct conversion of image sensor data to image analytics

Inventor: Pavel Sinha (Brossard, CA)
Assignee: Aarish Technologies
G06N3/084G06F18/2148G06F18/2155G06F18/24G06N3/04G06N3/08G06V10/147G06V10/454G06V10/82G06V10/98H04N23/84H04N25/13
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,198,059
App. No.
18/207,788
Granted
Jan 14, 2025
Kind
B2
Abstract

Systems and methods for performing direct conversion of image sensor data to image analytics are provided. One such system for directly processing sensor image data includes a sensor configured to capture an image and generate corresponding image data in a raw Bayer format, and a convolution neural network (CNN) coupled to the sensor and configured to generate image analytics directly from the image data in the raw Bayer format. Systems and methods for training the CNN are provided, and may include a generative model that is configured to convert RGB images into estimated images in the raw Bayer format.

Claims (45)

1. A system for directly processing sensor image data, the system comprising:

a sensor configured to capture an image and generate corresponding image data in a raw Bayer format; and

a convolution neural network (CNN) coupled to the sensor and configured to generate image analytics directly from the image data in the raw Bayer format,

wherein the CNN is configured to perform at least one of image classification or objection detection directly from the image data in the raw Bayer format.

2. The system of claim 1 :

wherein the CNN was trained using a generative model configured to convert an RGB image into an estimated image in the raw Bayer format using a ground truth image in the raw Bayer format; and

wherein the generative model was trained using image data in the raw Bayer format and without using labels.

3. The system of claim 2 , wherein the generative model is configured to generate a labeled image dataset in the raw Bayer format.

4. The system of claim 3 , wherein the CNN was trained using the labeled image dataset in the raw Bayer format.

5. A method for directly processing sensor image data, the method comprising:

receiving image data in a raw Bayer format;

generating image analytics directly from the image data in the raw Bayer format; and

performing at least one of image classification or objection detection directly from the image data in the raw Bayer format.

6. An apparatus for directly processing sensor image data, the apparatus comprising:

a means for receiving image data in a raw Bayer format;

a means for generating image analytics directly from the image data in the raw Bayer format; and

means for performing at least one of image classification or objection detection directly from the image data in the raw Bayer format.

7. A system for directly processing sensor image data, the system comprising:

a sensor configured to capture an image and generate corresponding image data in a raw RGB format;

a convolution neural network (CNN) coupled to the sensor and configured to generate image analytics directly from the image data in the raw RGB format; and

performing at least one of image classification or objection detection directly from the image data in the raw RGB format.

8. A system for directly processing sensor image data, the system comprising:

a sensor configured to capture an image and generate corresponding image data in a raw Bayer format; and

a convolution neural network (CNN) coupled to the sensor and configured to generate image analytics directly from the image data in the raw Bayer format,

wherein the CNN was trained using a generative model configured to convert an RGB image into an estimated image in the raw Bayer format using a ground truth image in the raw Bayer format.

9. A system for directly processing sensor image data, the system comprising:

a sensor configured to capture an image and generate corresponding image data in a raw Bayer format; and

a convolution neural network (CNN) coupled to the sensor and configured to generate image analytics directly from the image data in the raw Bayer format and without performing image signal pre-processing on the image data in the raw Bayer format.

10. A system for directly processing sensor image data, the system comprising:

a sensor configured to capture an image and generate corresponding image data in a raw Bayer format; and

a convolution neural network (CNN) coupled to the sensor and configured to generate image analytics directly from the image data in the raw Bayer format,

wherein the CNN is configured to generate inferences directly from the image data in the raw Bayer format.

11. The system of claim 1 , wherein the CNN was trained using a generative model configured to convert an RGB image into an estimated image in the raw Bayer format using a ground truth image in the raw Bayer format.

12. The system of claim 1 , wherein the CNN was trained using images in the raw Bayer format.

13. The system of claim 1 , wherein the CNN was trained using a labeled image dataset in the raw Bayer format.

14. The system of claim 10 , wherein the CNN was trained using a generative model configured to convert an RGB image into an estimated image in the raw Bayer format using a ground truth image in the raw Bayer format.

15. The system of claim 10 , wherein the CNN was trained using images in the raw Bayer format.

16. The system of claim 10 , wherein the CNN was trained using a labeled image dataset in the raw Bayer format.

17. A method for directly processing sensor image data, the method comprising:

receiving image data in a raw Bayer format;

generating, using a convolutional neural network (CNN), image analytics directly from the image data in the raw Bayer format; and

generating, using the CNN, inferences directly from the image data in the raw Bayer format.

18. The method of claim 17 , wherein the generating image analytics and the generating inferences is performed by the CNN trained using a generative model configured to convert an RGB image into an estimated image in the raw Bayer format using a ground truth image in the raw Bayer format.

19. The method of claim 17 , wherein the generating image analytics and the generating inferences is performed by the CNN trained using images in the raw Bayer format.

20. The method of claim 17 , wherein the generating image analytics and the generating inferences is performed by the CNN trained using a labeled image dataset in the raw Bayer format.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Sep 25, 2023
From: SINHA, PAVEL
To: AARISH TECHNOLOGIES
Reel/Frame 065011/0500 →
Continuity (4)
Continuation 17105293 · Nov 25, 2020
Provisional Application 63025580 · May 15, 2020
Provisional Application 62941646 · Nov 27, 2019
Related Publication 20230394314A1 · Dec 7, 2023
References Cited (19)
US 11158055B2 · Lin · 2021 [cited by examiner]
US 11301689B1 · Dreesch · 2022 [cited by examiner]
US 20180315155A1 · Park · 2018 [cited by examiner]
US 20200051217A1 · Shen et al. · 2020 [cited by applicant]
US 20200111198A1 · Hiasa · 2020 [cited by examiner]
US 20210035304A1 · Jie · 2021 [cited by examiner]
US 20210149005A1 · Wang · 2021 [cited by examiner]
US 20220318956A1 · Xu · 2022 [cited by examiner]
US 20220414849A1 · Wang · 2022 [cited by examiner]
WO 2018200075A1 · 2018 [cited by applicant]
Zhao, Yuzhi, et al. “Saliency map-aided generative adversarial network for raw to rgb mapping.” 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW). IEEE, 2019. (Year: 2019). [cited by examiner]
Jiang, Haomiao et al., “Learning the Image Processing Pipeline”, Cornell University; Computer Science: Computer Vision and Pattern Recognition; May 30, 2016; https://arxiv.org/pdf/1605.09336.pdf; 10 pages. [cited by applicant]
Barron, Jonathan T., “Convolutional Color Constancy”, 2015 IEEE International Conference on Computer Vision; Dec. 7-13, 2015; https://ieeexplore.ieee.org/document/7410408; 17 pages. [cited by applicant]
Bianco, Simone et al., “Color Constancy Using CNNs”, 2015 IEEE Conference on Computer Vision and Pattern Recognition Workshops; Jun. 7-12, 2015; https://ieeexplore.ieee.org/document/7301275; 9 pages. [cited by applicant]
Lou, Zhongyu et al., “Color Constancy by Deep Learning”, British Machine Vision Conference; Sep. 2015; https://ivi.fnwi.uva.nl/isis/publications/2015/LouBMVC2015/LouBMVC2015.pdf; 12 pages. [cited by applicant]
Qian, Yanlin et al., “Recurrent Color Constancy”, 2017 IEEE International Conference on Computer Vision; Oct. 22-29, 2017; https://ieeexplore.ieee.org/document/8237844; 9 pages. [cited by applicant]
Ratnasingam, Sivalogeswaran, “Deep Camera: A Fully Convolutional Neural Network for Image Signal Processing”, Cornell University; Electrical Engineering and Systems Science: Image and Video Processing; Aug. 24, 2019; ht… [cited by applicant]
PCT/IB2020/001007. International Search Report & Written Opinion (Mar. 25, 2021) pp. 8. [cited by applicant]
Krizhevsky et al., “ImageNet Classification with Deep Convolutional Neural Networks”, Communications of the ACM, Association for Computing Machinery, Inc., vol. 60, No. 6, 2012, pp. 84-90. <https:llpapers.nips.cc/paper/… [cited by applicant]