Human body tracing method, apparatus and device, and storage medium

Patent metadata

Field	Value
Publication number	US-11348354-B2
Application number	US-201916458531-A
Country	US
Kind code	B2
Filing date	Jul 1, 2019
Priority date	Jul 2, 2018
Publication date	May 31, 2022
Grant date	May 31, 2022

Abstract

A human body tracking method, apparatus, and device, and a storage medium. The method includes: obtaining a current frame image captured by a target photographing device at a current moment; detecting each human body in the current frame image to obtain first position information of the each human body in the current frame image; calculating second position information of a first human body in the current frame image; determining target position information of the each human body in the current frame image according to the second position information of the first human body in the current frame image, the first position information of the each human body in the current frame image, and pedestrian features of all tracked pedestrians stored in a preset list.

First claim

What is claimed is:

1. A human body tracking method, comprising:
obtaining a current frame image captured by a target photographing device at a current moment;
detecting each human body in the current frame image to obtain first position information of the each human body in the current frame image;
calculating, by using a kernelized correlation filters (KCF) tracking algorithm, second position information of a first human body in the current frame image, wherein the first human body is at least one human body tracked using the KCF tracking algorithm in a previous frame image ahead of the current frame image;
determining target position information of the each human body in the current frame image using the second position information of the first human body in the current frame image, the first position information of the each human body in the current frame image, and pedestrian features of all tracked pedestrians stored in a preset list; wherein the target position information is the more accurate one of the first position information and the second position information;
wherein the pedestrian features of all the tracked pedestrians stored in the preset list are image features in historical images captured by at least one photographing device at different historical moments.

2. The method according to claim 1 , wherein the determining target position information of the each human body in the current frame image using the second position information of the first human body in the current frame image, the first position information of the each human body in the current frame image, and pedestrian features of all tracked pedestrians stored in a preset list comprises:
determining target position information of the first human body in the current frame image using the second position information of the first human body in the current frame image, the first position information of the each human body in the current frame image, and the pedestrian features of all the tracked pedestrians stored in the preset list;
determining a second human body corresponding to a piece of first position information which does not match the second position information in the current frame image, according to the second position information of the first human body in the current frame image and the first position information of the each human body in the current frame image; and
determining target position information of the second human body in the current frame image using the piece of first position information of the second human body in the current frame image and the pedestrian features of all the tracked pedestrians stored in the preset list.

3. The method according to claim 2 , wherein the determining target position information of the first human body in the current frame image using the second position information of the first human body in the current frame image, the first position information of the each human body in the current frame image, and the pedestrian features of all the tracked pedestrians stored in the preset list comprises:
comparing the second position information of the first human body in the current frame image with the first position information of the each human body in the current frame image;
extracting an image feature of a first region corresponding to the second position information in the current frame image if the first position information of the each human body in the current frame image does not match the second position information;
comparing the image feature of the first region corresponding to the second position information in the current frame image with the pedestrian features of all the tracked pedestrians stored in the preset list; and
taking the second position information as the target position information of the first human body in the current frame image if a pedestrian feature that matches the image feature of the first region exists in the preset list.

4. The method according to claim 3 , wherein the determining target position information of the first human body in the current frame image using the second position information of the first human body in the current frame, the first position information of the each human body in the current frame image, and the pedestrian features of all the tracked pedestrians stored in the preset list further comprises:
in response to comparing the second position information of the first human body in the current frame image with the first position information of the each human body in the current frame image, taking the second position information as the target position information of the first human body in the current frame image if a piece of first position information that matches the second position information exists in the first position information of the each human body in the current frame image.

5. The method according to claim 2 , wherein the determining of target position information of the second human body in the current frame image according to the piece of first position information of the second human body in the current frame image and the pedestrian features of all the tracked pedestrians stored in the preset list comprises:
extracting an image feature of a second region corresponding to the second human body in the current frame image;
comparing the image feature of the second region with the pedestrian features of all the tracked pedestrians stored in the preset list; and
taking the piece of first position information of the second human body as the target position information of the second human body in the current frame image if a pedestrian feature that matches the image feature of the second region exists in the preset list.

6. The method according to claim 2 , wherein the method further comprises:
updating a tracker parameter corresponding to a preset tracking algorithm according to the target position information of the first human body in the current frame image and the target position information of the second human body in the current frame image.

7. The method according to claim 2 , wherein the method further comprises:
updating the pedestrian features of the tracked pedestrians in the preset list according to the target position information of the first human body in the current frame image and the target position information of the second human body in the current frame image.

8. An image processing device, comprising:
a memory;
a processor; and
a computer program;
wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method according to claim 1 .

9. The method of claim 1 , wherein it is performed by a non-transitory computer readable storage medium, wherein a computer program is stored thereon, and the computer program is executed by a processor.

10. A human body tracking apparatus, comprising: a processor and a computer-readable medium for storing program codes, which, when executed by the processor, cause the processor to: obtain a current frame image captured by a target photographing device at a current moment;
detect each human body in the current frame image to obtain first position information of the each human body in the current frame image;
calculate, by using a kernelized correlation filters (KCF), tracking algorithm, second position information of a first human body in the current frame image, wherein the first human body is at least one human body tracked using the KCF tracking algorithm in a previous frame image ahead of the current frame image; and
determine target position information of the each human body in the current frame image using the second position information of the first human body in the current fra

Assignees

• Baidu online network technology beijing co ltd

Inventors

• Wang Zhigang
• Wang Jian
• Li Xubin
• Kang Le
• Liu Zeyu
• Liu Xiao
• Sun Hao
• Wen Shilei
• Bao Yingze
• Chen Mingyu
• Ding Errui

Classifications

• G06V40/23

  Recognition of whole body movements, e.g. for sport training · CPC title

• G06V40/103 Primary

  Static body considered as a whole, e.g. static pedestrian or occupant recognition · CPC title

• G06V20/52 Primary

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

• G06F18/217

  Validation; Performance evaluation; Active pattern learning techniques · CPC title

• G06V40/20

  Movements or behaviour, e.g. gesture recognition (recognition of facial expressions G06V40/16) · CPC title