Research On The Methods Of Face Recognition For Mobile Phone

The technology based on face recognition is an important branch in the technology of biometric feature recognition because of its merits of non-contract,easy collection, it is now being widely used in the area of video surveillance,human computer interaction and so on.With the development of modern technology, online purchasing,mobile banking is more and more prevalent and has widely been accepted by people in our life. As an common communition tools in our daily life, mobile phone is now playing an important part like personal computer. Besides it has broad application prospects and space for development in many areas. While when we use the mobile phone the safety based on online payment is an important issue has to be considered. The text is done on the basis of the research of face recognition technology, by combining the face images gotten from the mobile phone to do the process of classification and recognition, accordingly we can achieve the purpose of improving the safety of mobile phone when using. The main task of this paper is as follows:1) Based on the need of the paper, constructe a multi-angle face database using the mobile phone as the precondition of the next work of face detection and face recognition;2) Make the summary and do the simulation of the main face detection methods now exsisting, by comparing the detection rate and the time complexity of them, an scheme which has better performance is given as the foundation of the subsequent face recognition methods;3) Considering the restriction of memory space and computing capability of mobile phone, the single sample face recognition method based on mirror Singular Value Decomposition is proposed, by comparing with the other single sample face recognition methods such as Principle Component Analysis based method and the Gabor method, in respect to the face recognition rate, the method proposed in this paper is improved more or less than the others.4) Based on the dimension reduction merits of Discrete Wavelet Decomposition and the ability of coefficient selection characteristics of Dynamic Weighted Discriminant Power Analysis(DWDPA), combining the two of them to the database coming from the mobile phone to do the task of face recognition, and give the comparition of the proposed method and the sub-space methods, the Dynamic Weighted Discriminant Power Analysis method and the Discriminant Power Analysis method on recognition rate and time complexity.5) Considering the application of the technology of face recognition on mobile phone, a system issue based on it is proposed, and the details of all parts of the system is analysed.