Face recognition algorithm based on Prewitt and convolutional neural network

Abstract

In order to improve the recognition performance of face recognition algorithm, a convolutional neural network face recognition method based on Prewitt operator is proposed. Firstly, the face image is preprocessed by histogram equalization and Prewitt operator. Then, the preprocessed images are input into the convolutional neural network for training, and exponential decay method is adopted to set the learning rate to accelerate the convergence rate, and L2 regulariz ation and Dropout are used to prevent overfitting. The recognition time of this method on ORL face database is 0.2s, and the recognition rate reaches 98.1%. The experimental results show that Prewitt operator and improved convolutional neural network can shorten the recognition time and improve the recognition rate.