A Convolutional Neural Network Face Recognition Algorithm Based on Data Augmentation

Changyu Cao1,a, Yang Wang2,b, Chih Cheng Chen3,c, Jiachun Zheng3,d
 
1Navigation Institute Jimei University Xiamen, Fujian
2Navigation Institute Jimei University Xiamen, Fujian
3Information Engineering College Jimei University Xiamen, Fujian
 
Abstract: 
Aiming at the problems of light changing and face blurring in face recognition based on convolution neural network,an optimization algorithm of data augmentationt is proposed, which uses LeNet and adds pool layer structure to extract face features.After the ORL database is enhanced, the training set and the test set are randomly generated for expeimental verification. The experimental results show that the accuracy of face recognition can reach 100% on the test set, and it has good robustness for some degree of illumination changes, facial expression changes and face blur.
 
Keywords: 

Convolutional neural network, face recognition, data augmentation

pages: 

147-150

DOI: 

10.35745/icice2018v2.037

Year: 

2018

Published in: 

2nd International Conference on Information, Communication and Engineering (ICICE 2018)

Full paper download: 
PDF icon 147-150.pdf