Systems and methods for object detection

Patent metadata

Field	Value
Publication number	US-10776665-B2
Application number	US-201815963340-A
Country	US
Kind code	B2
Filing date	Apr 26, 2018
Priority date	Apr 26, 2018
Publication date	Sep 15, 2020
Grant date	Sep 15, 2020

Abstract

A method performed by an electronic device is described. The method includes receiving a set of images. The method also includes determining a motion region and a static region based on the set of images. The method further includes extracting, at a first rate, first features from the motion region. The method additionally includes extracting, at a second rate that is different from the first rate, second features from the static region. The method also includes caching the second features. The method further includes detecting at least one object based on at least a portion of the first features.

First claim

What is claimed is:

1. A method performed by an electronic device, comprising:
receiving a set of images;
determining a motion region and a static region based on the set of images;
extracting, at a first rate, first features from the motion region;
extracting, at a second rate that is different from the first rate, second features from the static region, wherein extracting the first features and extracting the second features are performed on a union of the motion region and one or more regions of interest (ROIs) in the static region;
caching the second features; and
detecting at least one object based on at least a portion of the first features.

2. The method of claim 1 , further comprising:
detecting movement in the static region; and
retrieving information from a cache in response to the detected movement.

3. The method of claim 2 , wherein the information comprises cached features, and wherein the method further comprises:
determining a region of interest (ROI) based on the cached features; and
identifying an object based on the ROI and the cached features.

4. The method of claim 2 , wherein the information comprises a label, and wherein the method further comprises presenting the label.

5. The method of claim 1 , wherein the second rate is lower than the first rate.

6. The method of claim 1 , wherein a first operation thread that operates at the first rate comprises extracting the first features and detecting the at least one object based on the at least a portion of the first features, and a second operation thread that operates at the second rate comprises extracting the second features and caching the second features.

7. The method of claim 6 , wherein the second operation thread further comprises:
determining at least one region of interest (ROI) in the static region; and
detecting at least one object based on at least a portion of the second features in the at least one ROI.

8. The method of claim 7 , wherein the at least one ROI comprises a set of ROIs in the static region, and wherein extracting the second features comprises extracting features from at most a subset of the set of ROIs for each image of the set of images.

9. The method of claim 1 , wherein the first features and the second features are cached in a shared feature map.

10. The method of claim 1 , further comprising:
classifying the at least one object to produce at least one label; and
presenting the at least one label.

11. An electronic device, comprising:
a processor configured to:
receive a set of images;
determine a motion region and a static region based on the set of images;
extract, at a first rate, first features from the motion region;
extract, at a second rate that is different from the first rate, second features from the static region, wherein extracting the first features and extracting the second features are performed on a union of the motion region and one or more regions of interest (ROIs) in the static region;
cache the second features; and
detect at least one object based on at least a portion of the first features.

12. The electronic device of claim 11 , wherein the processor is configured to:
detect movement in the static region; and
retrieve information from a cache in response to the detected movement.

13. The electronic device of claim 12 , wherein the information comprises cached features, and wherein the processor is configured to:
determine a region of interest (ROI) based on the cached features; and
identify an object based on the ROI and the cached features.

14. The electronic device of claim 12 , wherein the information comprises a label, and wherein the processor is configured to present the label.

15. The electronic device of claim 11 , wherein the second rate is lower than the first rate.

16. The electronic device of claim 11 , wherein a first operation thread that operates at the first rate comprises extracting the first features and detecting the at least one object based on the at least a portion of the first features, and a second operation thread that operates at the second rate comprises extracting the second features and caching the second features.

17. The electronic device of claim 16 , wherein the second operation thread further comprises:
determining at least one region of interest (ROI) in the static region; and
detecting at least one object based on at least a portion of the second features in the at least one ROI.

18. The electronic device of claim 17 , wherein the at least one ROI comprises a set of ROIs in the static region, and wherein extracting the second features comprises extracting features from at most a subset of the set of ROIs for each image of the set of images.

19. The electronic device of claim 11 , wherein the first features and the second features are cached in a shared feature map.

20. The electronic device of claim 11 , wherein the processor is configured to:
classify the at least one object to produce at least one label; and
present the at least one label.

21. An apparatus, comprising:
means for receiving a set of images;
means for determining a motion region and a static region based on the set of images;
means for extracting, at a first rate, first features from the motion region;
means for extracting, at a second rate that is different from the first rate, second features from the static region, wherein the means for extracting the first features and the means for extracting the second features are based on a union of the motion region and one or more regions of interest (ROIs) in the static region;
means for caching the second features; and
means for detecting at least one object based on at least a portion of the first features.

22. The apparatus of claim 21 , further comprising:
means for detecting movement in the static region; and
means for retrieving information from a cache in response to the detected movement.

23. The apparatus of claim 21 , wherein the second rate is lower than the first rate.

24. A non-transitory tangible computer-readable medium storing computer executable code, comprising:
code for causing an electronic device to receive a set of images;
code for causing the electronic device to determine a motion region and a static region based on the set of images;
code for causing the electronic device to extract, at a first rate, first features from the motion region;
code for causing the electronic device to extract, at a second rate that is different from the first rate, second features from the static region, wherein the code for causing the electronic device to extract the first features and the code for causing the electronic device to extract the second features are based on a union of the motion region and one or more regions of interest (ROIs) in the static region;
code for causing the electronic device to cache the second features; and
code for causing the electronic device to detect at least one object based on at least a portion of the first features.

25. The computer-readable medium of claim 24 , further comprising:
code for causing the electronic device to detect movement in the static region; and
code for causing the electronic device to retrieve information from a cache in response to the detected movement.

26. The computer-readable medium of claim 24 , wherein the second rate is lower than the first rate.

Assignees

• Qualcomm Inc

Inventors

• Chen Ching-Hui
• Chiu Chinchuan
• Ma Lei
• Quan Shuxue

Classifications

• G06T7/215 Primary

  Motion-based segmentation · CPC title

• G06V20/52

  Surveillance or monitoring of activities, e.g. for recognising suspicious objects (recognising microscopic objects G06V20/69) · 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