Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Human face detecting and tracking method based on deep learning

A technology of face detection and deep learning, applied in the field of face detection and tracking based on deep learning, can solve problems such as information distortion and incorrect face comparison, achieve high information content and reduce repeated reporting

Active Publication Date: 2018-03-09
武汉众智数字技术有限公司
View PDF4 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, in the prior art, after the face feature points are extracted, the face obtained by rotating the face will distort the information, resulting in incorrect face comparison.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Human face detecting and tracking method based on deep learning
  • Human face detecting and tracking method based on deep learning
  • Human face detecting and tracking method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0035] Below in conjunction with accompanying drawing embodiment describe in detail:

[0036] 1. Method

[0037] Such as figure 1 , this method includes the following steps:

[0038] ①Face detection-101

[0039] Perform the face detection task of the frame of image data, and form a queue output for the detected target information, including the face rectangle frame, face category and face feature points;

[0040] ②Face Tracking-102

[0041] Match and update the result queue obtained by face detection to complete the task of face tracking;

[0042] ③Face judgment-103

[0043] Perform face frontal judgment on the latest tracked face in the track list, and output it as a detected high-quality face if certain conditions are met.

[0044] The specific implementation process of each module is described below:

[0045] 1. The process of step ①

[0046] Such as figure 2 , the process of step ① is as follows:

[0047] A. Use the P-net model trained offline for detection-201 ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a human face detecting and tracking method based on deep learning and relates to the technical field of video image processing. The method comprises the following steps: firstly, carrying out human face detection, namely carrying out a human face detection task of frame image data, and forming queue output for the detected target information comprising a human face rectangular frame, a human face type and human face feature points; secondly, carrying out human face tracking, namely matching and updating the result queue obtained after human face detection, and completing a human face tracking task; and thirdly, carrying out human face determination, namely carrying out front human face determination on the newest tracked human face in a tracking list, and outputtingas a detected high-quality human face when a certain condition is met. The method disclosed by the invention has the advantages that the problem that tracking is unsuccessful due to blockage and rotation of human face is comprehensively solved, and the problem of repeatedly reporting human face image data is reduced; a distinguishing method of achieving the aim of uploading by judging whether human face is a front human face or not is designed and trained, and the expected effect is achieved; and a reliable human face image with high information content is provided for algorithms such as human face matching.

Description

technical field [0001] The present invention relates to the technical field of video image processing, in particular to a face detection and tracking method based on deep learning. Background technique [0002] In the traditional face detection and tracking method, face detection is only good for the frontal face, not good for side faces or partial occlusion, and is affected by the intensity of light; in the process of face tracking, the face is easy to If it is lost due to occlusion or turning around, the face tracking will be terminated, and it will be tracked again as a new target after re-detection, causing the face to be captured repeatedly. [0003] There are restrictions on face comparison in the face recognition system, that is, the captured face picture is a positive face, and the comparison effect will be better. However, in the prior art, after the face feature points are extracted, the face obtained by rotating the face will distort the information, resulting in...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06K9/00
CPCG06V40/161G06V40/168
Inventor 陆辉刘树惠
Owner 武汉众智数字技术有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com