Research And Application Of Face Recognition Based On Artificial Neural Network

With the continuous progress of the society and the rapid development of the computer technology, the traditional authentication technology has been unable to meet the needs of the development of modern science and human society. Compared with other identification methods, the nature of the facial recognition is more friendly and easy to accept, and it is wide research and application. Therefore, a method of face image recognition and image feature extraction based on artificial neural network is put forward. In this method for distinguishing, the first step is to use block PCA algorithm for the face image processing, and gain the face image features of different block number. These features are the inputs of artificial neural networks, the output is the corresponding category.The experiment is based on ORL face images database and partitioning PCA algorithm, the simulation partition of the face images into the sizes of 1*1, 2*1, 2*2, 4*2 and 4*4 for the testing respectively. The results show that with the increase of face image block number, the complexity of the artificial neural network also will increase, so we need many training steps to complete the training of artificial neural network. With the increase of face image block number, the complexity of the neural network structure will increase also, its ability of the pattern recognition will also increase. So with the complete training of the artificial neural network, the recognition accuracy is become higher. Through the above experiment results which show that face image recognition based on the block PCA algorithm and artificial neural network is effective. The application of block PCA algorithm improves the identification accuracy of face image recognition. In the future, the research will continue to study the block number and to improve the testing precision.