Background: Face recognition has a very important application value in the field of information
security as an important method of bioinformatics identification. There are recent patents
that discuss a human face similarity recognition method and system. It has also faced the problem of
complex feature space and the very large amount of data, which make face recognition one of the
most challenging and most academic research topics.
Method: In order to solve the problem of the lack of prior knowledge in the face recognition algorithm
based on the traditional convolution neural network, this paper improves the traditional convolution
neural network from the two aspects of feature extraction and classification recognition, and
proposes a new method-face recognition algorithm based on the sparse representation of denoising
autoencoder convolution neural network, SRDAECNN.
Results: Extensive experiments are performed on LFW, ORL, YALE and other face database. The
experimental results show that our proposed face recognition algorithm has high accuracy.
Conclusion: The model combines the advantages of the convolution neural network and sparse representation-
based classifier, which can overcome the problem of incomplete feature extraction due
to the random initialization of convolution kernel, and introduce sparse representation algorithm on
classification recognition to enhance the recognition effect.