The Research On Human Face Recognition Based On Artificial Neural Network

Human face image recognition is a kind of facial features as identification ofbiological recognition technology.In the digital information age, along with the moreand more strict identification, and the development of the digital image identificationtechnology, face recognition technology is more and more get the favour of people. Atpresent, face recognition technology are mainly based on analysis of characteristicsand based on the overall analysis of the two kinds of methods.Face recognitioninvolves mathematics, biology, computer vision and brain sciences, many disciplinessuch as research methods are complex and changeable. This article through to study inthe field of face recognition, and carried out exploratory researchIn the face detection phase, through the analysis of characteristics of face, todesign a detection model based on Adaboost algorithm and BP neural network.In viewof the human face image preprocessing, using switch strategy combined with PCNN adenoising algorithm is designed, good protection the texture of the face.In the featureextraction stage, through the fiducial points setting makes PCA feature classificationmore accurate. Finally, in terms of face feature matching, through the combination ofPCA and BPNN, compared to pure PCA matching to improve the accuracy.