Research on encrypted face recognition algorithm based on new combined chaotic map and neural network

Abstract

Face recognition technology is currently one of the most popular and widely used artificial intelligence technologies. The application system based on face recognition technology brings many new security and privacy issues. The face image database of the traditional face recognition method is stored in the cloud, which may be attacked by hackers and may lead to the leakage of personal privacy and data security problems of users. Aiming at the security problem of traditional face recognition, this paper proposes an encrypted face recognition algorithm based on a new combined chaotic map and neural network. Firstly, this method uses Logistic and Sine chaotic mapping to generate keys, then encrypt the face database image. Next, combine the PCA method and neural network to realize encrypted face recognition, and finally perform conventional attacks, geometric attacks and occlusion attacks to test its robustness. The experimental results show that the recognition rate of the encryption algorithm is 92.5%, which is higher than the recognition rate of the unencrypted algorithm by 11.5%. This indicates that it has good encryption effect and robustness while guaranteeing the recognition rate and also has a good application prospect.