Abstract: In order to solve the problem of low recognition rate of face recognition algorithm in face images with different pose，an improved algorithm based on convolutional neural network is proposed． The algorithm optimize the classical convolutional neural network LeNet-5's convolution kernel number，convolution kernel size，pool layer mode，regularization method，The pose-varied Face's feature can be extracted effectively． The experimental results show that the algorithm improves the rate and recognition time significantly．

Key words: face recognition, convolutional neural network, unconstrained environment, image processing