IP Library Granted Patent US 11,676,367
Granted Patent B1
US 11,676,367 · App. 16/684,458 · Granted Jun 13, 2023

System and method for anomaly detection using anomaly cueing

Inventor: Yuri Owechko (Newbury Park, CA)
Assignee: HRL LABORATORIES, LLC
G06V10/751G06T7/62G06T7/90H03H21/0025G06F18/23G06F2218/08H03H2021/0034
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 11,676,367
App. No.
16/684,458
Granted
Jun 13, 2023
Kind
B1
Abstract

Described a system for anomaly detection using anomaly cueing. In operation, an input image having two-dimensional (2D) image mixtures of primary components is reformatted into one-dimensional (1D) input signals. Blind source signal separation is used to separate the 1D input signals into separate output primary components, which are 1D output signals. The 1D output signals are reformatted into 2D spatially independent component output images. The system then calculates all possible pair product images of the 2D spatially independent component output images and corresponding signal-to-noise ratios. A pair product image is selected based on the peak signal-to-noise ratio and thresholded to identify anomalies in the pair product image. Several types of devices can then be controlled based on the identified anomalies in the pair product image.

Claims (48)

1. A system for anomaly detection using anomaly cueing, the system comprising:

one or more processors and a memory, the memory being a non-transitory computer-readable medium having executable instructions encoded thereon, such that upon execution of the instructions, the one or more processors perform operations of:

receiving an input image having two-dimensional (2D) image mixtures of primary components;

reformatting the 2D image mixtures into one-dimensional (1D) input signals;

using blind source signal separation, separating the 1D input signals into separate output primary components, the separate output primary components being 1D output signals;

reforming the 1D output signals into 2D spatially independent component output images

calculating all possible pair product images of the 2D spatially independent component output images and corresponding signal-to-noise ratios;

selecting a pair product image based on the signal-to-noise ratio; and

thresholding the selected pair product image to identify anomalies in the pair product image.

2. The system as set forth in claim 1 , further comprising an operation of controlling a device based on the anomalies.

3. The system as set forth in claim 1 , further comprising an operation of monitoring image statistics of a next image frame to determine if the 2D image mixtures of primary components have changed from those of the input image, such that if the 2D image mixtures of primary components have not changed, then reusing a demixing matrix from a previous frame to demix the next image frame, otherwise performing anomaly detection on the next image frame to identify anomalies on the next image frame.

4. The system as set forth in claim 1 , further comprising an operation of using a subset of the input image to calculate a demixing matrix and then using the demixing matrix to demix the input image based on the subset of the input image.

5. The system as set forth in claim 1 , wherein the 2D image mixtures are formed of color, spectral, or polarimetric channels of image data.

6. The system as set forth in claim 1 , wherein the blind source signal separation is independent component analysis.

7. The system as set forth in claim 1 , wherein the 2D image mixtures are reformatted into 1D input signals by concatenating rows and treating the 1D input signal as a mixture of image components for input to the blind source signal separation.

8. The system as set forth in claim 1 , where the pair product images having dimensions matching the input image such that locations of anomalies in the pair product images correspond directly to locations in the input image.

9. A computer implemented method for anomaly detection using anomaly cueing, the method comprising an act of:

causing one or more processers to execute instructions encoded on a non-transitory computer-readable medium, such that upon execution, the one or more processors perform operations of:

receiving an input image having two-dimensional (2D) image mixtures of primary components;

reformatting the 2D image mixtures into one-dimensional (1D) input signals;

using blind source signal separation, separating the 1D input signals into separate output primary components, the separate output primary components being 1D output signals;

reforming the 1D output signals into 2D spatially independent component output images;

calculating all possible pair product images of the 2D spatially independent component output images and corresponding signal-to-noise ratios;

selecting a pair product image based on the signal-to-noise ratio; and

thresholding the selected pair product image to identify anomalies in the pair product image.

10. The method as set forth in claim 9 , further comprising an operation of controlling a device based on the anomalies.

11. The method as set forth in claim 9 , further comprising an operation of monitoring image statistics of a next image frame to determine if the 2D image mixtures of primary components have changed from those of the input image, such that if the 2D image mixtures of primary components have not changed, then reusing a demixing matrix from a previous frame to demix the next image frame, otherwise performing anomaly detection on the next image frame to identify anomalies on the next image frame.

12. The method as set forth in claim 9 , further comprising an operation of using a subset of the input image to calculate a demixing matrix and then using the demixing matrix to demix the input image based on the subset of the input image.

13. The method as set forth in claim 9 , wherein the 2D image mixtures are formed of color, spectral, or polarimetric channels of image data.

14. The method as set forth in claim 9 , wherein the blind source signal separation is independent component analysis.

15. The method as set forth in claim 9 , wherein the 2D image mixtures are reformatted into 1D input signals by concatenating rows and treating the 1D input signal as a mixture of image components for input to the blind source signal separation.

16. The method as set forth in claim 9 , where the pair product images having dimensions matching the input image such that locations of anomalies in the pair product images correspond directly to locations in the input image.

17. A computer program product for anomaly detection using anomaly cueing, the computer program product comprising:

a non-transitory computer-readable medium having executable instructions encoded thereon, such that upon execution of the instructions by one or more processors, the one or more processors perform operations of:

receiving an input image having two-dimensional (2D) image mixtures of primary components;

reformatting the 2D image mixtures into one-dimensional (1D) input signals;

using blind source signal separation, separating the 1D input signals into separate output primary components, the separate output primary components being 1D output signals;

reforming the 1D output signals into 2D spatially independent component output images;

calculating all possible pair product images of the 2D spatially independent component output images and corresponding signal-to-noise ratios;

selecting a pair product image based on the signal-to-noise ratio; and

thresholding the selected pair product image to identify anomalies in the pair product image.

18. The computer program product as set forth in claim 17 , further comprising an operation of controlling a device based on the anomalies.

19. The computer program product as set forth in claim 17 , further comprising an operation of monitoring image statistics of a next image frame to determine if the 2D image mixtures of primary components have changed from those of the input image, such that if the 2D image mixtures of primary components have not changed, then reusing a demixing matrix from a previous frame to demix the next image frame, otherwise performing anomaly detection on the next image frame to identify anomalies on the next image frame.

20. The computer program product as set forth in claim 17 , further comprising an operation of using a subset of the input image to calculate a demixing matrix and then using the demixing matrix to demix the input image based on the subset of the input image.

21. The computer program product as set forth in claim 17 , wherein the 2D image mixtures are formed of color, spectral, or polarimetric channels of image data.

22. The computer program product as set forth in claim 17 , wherein the blind source signal separation is independent component analysis.

23. The computer program product as set forth in claim 17 , wherein the 2D image mixtures are reformatted into 1D input signals by concatenating rows and treating the 1D input signal as a mixture of image components for input to the blind source signal separation.

24. The computer program product as set forth in claim 17 , where the pair product images having dimensions matching the input image such that locations of anomalies in the pair product images correspond directly to locations in the input image.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Nov 14, 2019
From: OWECHKO, YURI
To: HRL LABORATORIES, LLC
Reel/Frame 051020/0724 →
Continuity (1)
Provisional Application 62802438 · Feb 7, 2019
Cited By (2)
US 12,353,977 US 12,443,501