A Face Recognition Method Based on DiaPCA and Neural Network

Putting forward a face recognition method based on Diagonal Principal Component Analysis and BP neural network. Firstly, do the dimension reduction to the sample data and take the DiaPCA method to avoid the information drop; Then, use the classics BP neural network to do the face detection. It not only shorten the net training time, but also improve the accuracy of the recognition. It used 1000 face and nonface images of the MITEx face database to test and analysis, compared with the method that combined LLE and neural network. The simulation demonstrates that its effect of detection is better.