Method and apparatus for face detection in a frame sequence using sub-tasks and layers

Patent metadata

Field	Value
Publication number	US-9965673-B2
Application number	US-201113977684-A
Country	US
Kind code	B2
Filing date	Apr 11, 2011
Priority date	Apr 11, 2011
Publication date	May 8, 2018
Grant date	May 8, 2018

Abstract

Techniques are disclosed that involve face detection. For instance, face detection tasks may be decomposed into sets of one or more sub-tasks. In turn the sub-tasks of the sets may be allocated across multiple image frames. This allocation may be based on a multiple layer, quad-tree approach. In addition, face tracking tasks may be performed.

First claim

The invention claimed is:

1. A method of performing a face detection task as a plurality of sub-tasks, comprising:
receiving a plurality of image frames in a video stream sequence of frames; and
performing a first face detection sub-task on a single detection window at a first layer on a first one of the sequence of frames, the first layer having a first detection window size range corresponding to a first size of a portion of the first frame;
performing a second face detection sub-task on a single detection window at a second layer different from the first layer on a second one of the sequence of frames subsequent to the first one of the sequence of frames, the second layer having a second detection window size range corresponding to a second size of a portion of the second frame different from the first size of the portion of the first one of the sequence of frames;
performing a third face detection sub-task on a single detection window at a third layer different from the first layer and from the second layer on a third one of the sequence of frames subsequent to the second one of the sequence of frames, the third layer having a third detection window size range corresponding to a third size of a portion of the third frame different from the first size and from the second size, wherein the sub-tasks are distributed across multiple frames by performing only one sub-task on each frame;
generating face detection data based on the performed face detection sub-tasks; and
sending the face detection data to a data management module.

2. The method of claim 1 , wherein said performing comprises, for each sub-task, scanning a window within a detection region, the window having a size within the size range of the corresponding layer.

3. The method of claim 2 , wherein the detection region has a size within the size range of the corresponding layer.

4. The method of claim 1 :
wherein the plurality of layers includes a first layer employing a first quantity of detection regions, and a second successive layer employing a second quantity of detection regions; and
wherein the second quantity is four times greater than the first quantity.

5. The method of claim 4 , wherein each detection region corresponds to a sub-task and wherein said performing a second face detection sub-task comprises performing the second face detection sub-task in a next frame of the sequence of frames followed by performing a sub-task for each other detection region at the second layer in sequential frames of the sequence of frames before performing the third face detection sub-task.

6. The method of claim 1 , further comprising:
performing a face tracking task on the plurality of image frames.

7. The method of claim 1 , wherein each face detection sub-task causes the detection of faces, the method further comprising maintaining information regarding one or more previously detected faces.

8. An apparatus, comprising:
a face detection module to receive a video stream sequence of frames, to perform a face detection task on a plurality of image frames, to generate face detection data and to send the face detection data to a data management module,
the face detection module performing a first face detection sub-task on a single detection window at a first layer on a first one of the sequence of frames, the first layer having a first detection window size range corresponding to a first size of a portion of the first frame,
performing a second face detection sub-task on a single detection window at a second layer different from the first layer on a second one of the sequence of frames subsequent to the first one of the sequence of frames, the second layer having a second detection window size range corresponding to a second size of a portion of the second frame different from the first size of the portion of the first one of the sequence of frames, and
performing a third face detection sub-task on a single detection window at a third layer different from the first layer and from the second layer on a third one of the sequence of frame subsequent to the second one of the sequence of frames, the third layer having a third detection window size range corresponding to a third size of a portion of the third frame different from the first size and from the second size,
wherein the sub-tasks are distributed across multiple frames by performing only one sub-task on each frame.

9. The apparatus of claim 8 , wherein said face detection module is to, for each sub-task, scan a window within a detection region, the window having a size within the size range of the corresponding layer.

10. The apparatus of claim 9 , wherein the detection region has a size within the size range of the corresponding layer.

11. The apparatus of claim 8 :
wherein the plurality of layers includes a first layer employing a first quantity of detection regions, and a second successive layer employing a second quantity of detection regions; and
wherein the second quantity is four times greater than the first quantity.

12. The apparatus of claim 11 , wherein each detection region corresponds to a sub-task and wherein said face detection module is to perform a second face detection sub-task in a next frame of the sequence of frames followed by performing a sub-task for each other detection region at the second layer in sequential frames of the sequence of frames before performing the third face detection sub-task.

13. The apparatus of claim 8 , further comprising a face tracking module to track one or more previously detected faces within an image frame.

14. The apparatus of claim 13 , further comprising a storage medium to store information regarding the one or more previously detected faces.

15. The apparatus of claim 14 , wherein the storage medium further stores information regarding any faces detected by one or more face detection sub-tasks.

16. The apparatus of claim 8 , further comprising an image sensor to generate a plurality of frames in the video stream sequence of frames.

17. The apparatus of claim 8 , further comprising a display to output the image frame.

18. An article comprising a non-transitory machine-accessible medium having stored thereon instructions that, when executed by a machine, cause the machine to:
receive a plurality of image frames in a video stream sequence of frames; and
perform a first face detection sub-task on a single detection window at a first layer on a first one in sequence of the sequence of frames, the first layer having a first detection window size range corresponding to a first size of a portion of the first frame, wherein the face detection task comprises a plurality of face detection sub-tasks, the plurality of sub-tasks allocated among a plurality of layers;
perform a second face detection sub-task on a single detection window at a second layer different from the first layer on a second one in sequence of the sequence of frames subsequent to the first one of the sequence of frames, the second layer having a second detection window size range corresponding to a second size of a portion of the second frame different from the first size of the portion of the first one of the sequence of frames; and
perform a third face detection sub-task on a single detection window at a third layer different from the first layer and from the second layer on a third one in sequence of the sequence of frames subsequent to the second one of the sequence of frames, the third layer having a third detection window size range corresponding to a third size of a portion of the third frame different from the fir

Assignees

• Du Yangzhou
• Li Jianguo
• Liu Ang
• Wang Tao
• Zhang Yimin
• Intel Corp

Inventors

• Du Yangzhou
• Li Jianguo
• Liu Ang
• Wang Tao
• Zhang Yimin

Classifications

• G06V40/167

  using comparisons between temporally consecutive images · CPC title

• G06T7/248 Primary

  involving reference images or patches · CPC title

• G06T2207/30201

  Face · CPC title

• G06T2207/20016

  Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform · CPC title

• G06K9/00248

  Physics · mapped topic