Visual anomaly detection without reference in graphics computing environments

Patent metadata

Field	Value
Publication number	US-11301686-B2
Application number	US-201916417981-A
Country	US
Kind code	B2
Filing date	May 21, 2019
Priority date	May 25, 2018
Publication date	Apr 12, 2022
Grant date	Apr 12, 2022

Abstract

A mechanism is described for facilitating visual anomaly detection without reference in computing environments. An apparatus of embodiments, as described herein, includes one or more processors to select a frame from a sequence of multiple frames associated with a video stream captured by a camera, and dynamically compute a frame confidence score for the frame based on frame training data associated with frame. The one or more processors are further to detect one or more anomalies in the frame when the frame confidence score is less than a frame confidence threshold associated with the frame, where detecting includes dynamically comparing the frame confidence score with the frame confidence threshold through inference using frame field data and the frame training data.

First claim

What is claimed is:

1. An apparatus comprising:
one or more processors to:
select a frame from a sequence of multiple frames associated with a video stream captured by a camera;
dynamically compute a frame confidence score and a sequence confidence score, wherein the frame confidence score is computed for the frame based on frame training data associated with the frame, and wherein the sequence confidence score is computed for the sequence based on sequence training data associated with the sequence; and
detect one or more anomalies in the frame when the frame confidence score is less than a frame confidence threshold associated with the frame, wherein detecting includes dynamically comparing the frame confidence score with the frame confidence threshold through inference based on frame field data and the frame training data.

2. The apparatus of claim 1 , wherein the one or more processors are further to reduce a frame size of the frame by splitting the frame into multiple frame portions, wherein the one or more anomalies are detected in one or more of the multiple frame portions.

3. The apparatus of claim 1 , wherein the one or more processors are further to select the sequence of multiple frames associated with the video stream.

4. The apparatus of claim 3 , wherein the one or more processors are further to:
detect one or more anomalies in the sequence when the sequence confidence score is less than a sequence confidence threshold associated with the sequence, wherein detecting includes dynamically comparing the sequence confidence score with the sequence confidence threshold through inference using sequence field data and the sequence training data.

5. The apparatus of claim 4 , wherein the one or more anomalies associated with the sequence are independent and irrespective of corruption status associated with each of the multiple frames of the sequence.

6. The apparatus of claim 4 , wherein the one or more processors are further to reduce a sequence size of the sequence by splitting one or more of the multiple frames into smaller frames.

7. The apparatus of claim 1 , wherein the one or more processors comprise a graphics processor hosting a no-reference anomaly detection circuitry, wherein the one or more processors further comprise an application processor co-located with the graphics processor on a common semiconductor package.

8. A method comprising:
selecting, by one or more processors, a frame from a sequence of multiple frames associated with a video stream captured by a camera;
dynamically computing, by the one or more processors, a frame confidence score and a sequence confidence score, wherein the frame confidence score is computed for the frame based on frame training data associated with the frame, and wherein the sequence confidence score is computed for the sequence based on sequence training data associated with the sequence; and
detecting, by the one or more processors, one or more anomalies in the frame when the frame confidence score is less than a frame confidence threshold associated with the frame, wherein detecting includes dynamically comparing the frame confidence score with the frame confidence threshold through inference based on frame field data and the frame training data.

9. The method of claim 8 , further comprising reducing a frame size of the frame by splitting the frame into multiple frame portions, wherein the one or more anomalies are detected in one or more of the multiple frame portions.

10. The method of claim 8 , further comprising selecting the sequence of multiple frames associated with the video stream.

11. The method of claim 10 , further comprising:
detecting one or more anomalies in the sequence when the sequence confidence score is less than a sequence confidence threshold associated with the sequence, wherein detecting includes dynamically comparing the sequence confidence score with the sequence confidence threshold through inference using sequence field data and the sequence training data.

12. The method of claim 11 , wherein the one or more anomalies associated with the sequence are independent and irrespective of corruption status associated with each of the multiple frames of the sequence.

13. The method of claim 11 , further comprising reducing a sequence size of the sequence by splitting one or more of the multiple frames into smaller frames.

14. The method of claim 8 , wherein the one or more processors comprise a graphics processor hosting a no-reference anomaly detection circuitry, wherein the one or more processors further comprise an application processor co-located with the graphics processor on a common semiconductor package.

15. At least one non-transitory machine-readable medium comprising a plurality of instructions which, when executed on a processing device, cause the processing device to perform operations comprising:
selecting a frame from a sequence of multiple frames associated with a video stream captured by a camera;
dynamically computing a frame confidence score and a sequence confidence score, wherein the frame confidence score is computed for the frame based on frame training data associated with the frame, and wherein the sequence confidence score is computed for the sequence based on sequence training data associated with the sequence; and
detecting one or more anomalies in the frame when the frame confidence score is less than a frame confidence threshold associated with the frame, wherein detecting includes dynamically comparing the frame confidence score with the frame confidence threshold through inference based on frame field data and the frame training data.

16. The non-transitory machine-readable medium of claim 15 , wherein the operations further comprise reducing a frame size of the frame by splitting the frame into multiple frame portions, wherein the one or more anomalies are detected in one or more of the multiple frame portions.

17. The non-transitory machine-readable medium of claim 15 , wherein the operations further comprise selecting the sequence of multiple frames associated with the video stream.

18. The non-transitory machine-readable medium of claim 17 , wherein the operations further comprise:
detecting one or more anomalies in the sequence when the sequence confidence score is less than a sequence confidence threshold associated with the sequence, wherein detecting includes dynamically comparing the sequence confidence score with the sequence confidence threshold through inference using sequence field data and the sequence training data.

19. The non-transitory machine-readable medium of claim 18 , wherein the one or more anomalies associated with the sequence are independent and irrespective of corruption status associated with each of the multiple frames of the sequence.

20. The non-transitory machine-readable medium of claim 18 , wherein the operations further comprise reducing a sequence size of the sequence by splitting one or more of the multiple frames into smaller frames.

21. The non-transitory machine-readable medium of claim 15 , wherein the processing device comprises a graphics processor hosting a no-reference anomaly detection circuitry, wherein the processing device further comprises an application processor co-located with the graphics processor on a common semiconductor package.

Assignees

• Intel Corp

Inventors

• Muppalla Ashwin
• Save Sanket
• Sudireddy Subash
• Armon Amitai
• Faivishevsky Lev
• Fania Moty
• Hollander Tahi

Classifications

• G06T7/0002 Primary

  Inspection of images, e.g. flaw detection · CPC title

• G06V10/993

  Evaluation of the quality of the acquired pattern · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/454

  Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN] · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title