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

Face identification method based on deep convolutional neural network

A neural network and deep convolution technology, applied in the field of face recognition based on deep convolutional neural network, can solve the problems of poor face recognition accuracy, limited application range, and excessive calculation, so as to avoid large databases The effect of dependence, improving accuracy and reducing complexity

Inactive Publication Date: 2017-11-07
TIANJIN UNIV
View PDF2 Cites 26 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The main challenges currently faced in the field of face recognition are: when training deep neural network structures, most methods are highly dependent on the capacity of the database, resulting in excessive calculation and the accuracy of face recognition is not very good, which limits the actual Application range

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
  • Face identification method based on deep convolutional neural network
  • Face identification method based on deep convolutional neural network
  • Face identification method based on deep convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 2

[0048] Combined with the specific calculation formula, figure 2 ,as well as image 3 The scheme in embodiment 1 is further introduced, see the following description for details:

[0049] 201: Use a multi-step processing strategy to build a small face recognition database from scratch;

[0050] Wherein, establishing a face recognition database includes the following steps:

[0051]1. First, get 5,000 names from the list of Chinese actors according to their popularity, half of them are male and half male. These names are eliminated through continuous screening, and finally the names of N (N=2622) individuals are obtained.

[0052] Wherein, the above-mentioned process of screening and exclusion can be adopted in any manner known to those skilled in the art, and there is no limit to the number of names to be selected.

[0053] 2. With the help of Baidu and Google image search engines, search according to "person's name" and "person's name and actor" respectively, and select t...

Embodiment 3

[0086] The following is combined with specific calculation formulas, examples, and Figure 4 The scheme in embodiment 1 and 2 is carried out feasibility verification, see the following description for details:

[0087] The database used in this experiment training is the face recognition database constructed in step 201. This is a face database about Chinese actors, which contains a total of 2,622 names, and each name corresponds to 375 pictures, a total of 983,250 pictures.

[0088] The database used in this experiment is Labeled Faces in the Wild dataset (LFW) [8] , which has become a standard database for evaluating recognition performance in academia, which contains 5749 identities with a total of 13233 images. The database is well known to those skilled in the art, and will not be described in detail in this embodiment of the present invention.

[0089] Without loss of generality, the average recognition accuracy To measure the recognition performance of this method....

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 present invention discloses a face identification method based on a deep convolutional neural network. The method comprises: respectively feeding each image in a face identification database into three constructed deep convolutional neural network for feature extraction; performing normalization of output features, performing affine projection of the features to a low-dimensional space, obtaining a projection matrix, training a projection matrix through minimization of a ternary loss function, and obtaining feature vectors of each image; searching the weight value of each filter in the deep convolutional neural network through a gradient descending method, performing training test, and selecting a deep convolutional neural network having the highest average identification precision; and applying the selected deep convolutional neural network to a standard face identification database, performing Euclidean distance calculation of feature vectors of face images to be detected and each image, and if the feature vectors are smaller than a threshold value, determining that the images show the same person. Few images are used in the training, and the employed convolutional neural network is simple in structure so as to improve the face identification precision and reduce the training complexity.

Description

technical field [0001] The invention relates to the field of face recognition, in particular to a face recognition method based on a deep convolutional neural network. Background technique [0002] In modern society, the application of personal identity authentication technology is ubiquitous, and the recognition technology based on fingerprint, iris, and human face and other human biometrics has huge market demand in many fields, such as: access control system, video surveillance, airport security check, and smart spaces. Although identity authentication based on fingerprints and irises has higher accuracy and reliability than face recognition technology, face recognition has a wider range of applications due to its advantages of being natural, friendly, less disturbing to users, and easy to be accepted by users. prospect [1] . [0003] Face recognition is a process of analyzing and comparing face images in the database based on technologies such as digital image process...

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 Applications(China)
IPC IPC(8): G06K9/00G06N3/04
CPCG06N3/04G06V40/172
Inventor 聂为之王洪涛
Owner TIANJIN UNIV
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