A fast face recognition method against angle and occlusion interference

A face recognition, fast technology, applied in the field of biometrics, can solve the problems of limited ability of face image connotation features, unsatisfactory recognition effect, etc., to avoid slow convergence speed, simple comparison method, and obvious effect Effect

Active Publication Date: 2022-03-25
SOUTH CHINA UNIV OF TECH
View PDF11 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This kind of method has limited ability to extract the connotative features of face images, and its recognition effect is also unsatisfactory.

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
  • A fast face recognition method against angle and occlusion interference
  • A fast face recognition method against angle and occlusion interference
  • A fast face recognition method against angle and occlusion interference

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0045] like figure 1 As shown, the main process of the fast face recognition method for resisting angle and occlusion interference in this embodiment includes the following steps:

[0046] The first step is to obtain the training set, expand, adjust and preprocess the training set.

[0047] Adjust the face training set to ensure that the profile and glasses-wearing face pictures account for 10% to 20% of the training set. Using random cropping and random flipping to expand the dataset, the single-channel images in the dataset are expanded into three-channel images, and the expanded training set is adjusted to a size of 160x160. The method of extending a three-channel image from a single-channel image is to assign the value of each pixel of the image to the RGB three-channel value of the pixel of the single-channel image. The image is normalized, and the normalization method is x is the image pixel value, image_x std is the normalized pixel value, mean is the average val...

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 provides a fast face recognition method that is resistant to angle and occlusion interference. Firstly, a training set of pictures that have been detected and cut out of faces is collected, the training set is adjusted and expanded, and standardized preprocessing is performed. Put into the constructed neural network for training and save the parameter model. Adjust the test data to an appropriate size and number of channels, and also perform standardized preprocessing, put it into the prediction network, and obtain the feature vector of the face picture, and judge whether two faces come from the same face by calculating the distance between the feature vectors of the face picture people. The present invention has no specific requirement on the angle of the face photo, and has obvious comparison effect on the face photos wearing glasses, so there is no need to worry about the influence of side face angle and glasses occlusion on the face comparison. In addition, the comparison method of the present invention is simple, uses GPU acceleration to improve computing performance, and can be applied to a real-time environment.

Description

technical field [0001] The invention belongs to the technical field of biometric identification, and relates to a fast face identification method which is resistant to angle and occlusion interference. Background technique [0002] The development of face recognition technology has emerged since the 1970s, and it is one of the hottest research problems in the field of computer vision so far. In recent years, with the rapid development of modern information technology and the urgent need for automatic identity verification, biometric identification technology has developed rapidly in recent years. Modern biometric technology is mainly through the combination of computer and advanced technical means, using the inherent physiological characteristics and behavioral characteristics of the human body to identify and identify personal identity. Mainly include: fingerprint identification, retinal identification, iris identification, vein identification, gait identification, face id...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06V40/16G06V10/74G06V10/764G06V10/772G06V10/82G06V10/94G06K9/62G06V10/774
CPCG06V40/16G06V40/168G06V40/172G06F18/22G06F18/24G06F18/214G06N3/04G06N3/063G06N3/082G06V10/454G06V10/82G06V10/764G06V10/774G06N3/048G06N3/08G06T3/4046G06T2207/20081G06T2207/20084G06T2207/20132G06T2207/30201G06V40/161G06F18/2413
Inventor 黄翰刘迪郝志峰
Owner SOUTH CHINA UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap