Statistical Characteristics Of Human Face Recognition System Based On Wavelet Transform

The technique of face recognition is an application of the image manipulation, it has very strong aptitude, it is a new identified system, the application of it is feasible, it bring convenience and safety, and it has high fact value to study the applications of it.Recognized technique with biological feature is attached importance to more and more with the development of information technology. Some mature products about face recognition come forth in recent years, but the recognized effect of the face recognized systems is not perfect with a lot of disturbance from the changes of the human face, so nowadays there are many scholars to do many of works in studying the technology. The solution of the general problem of face recognition contains three aspects:1. Face orientation: Go to orientation the face from cluttered scenes.2. Facial features detection: Location of the facial features from the face region.3. Face recognition: Extraction of features and recognition.The paper premises the front two aspects are solved and mostly studiesthe face recognition with wavelet transform. Firstly the face picture isanalyzed to high frequency part and low frequency part, second we coarselyclassify the face identified by the histogram of the wavelet low frequencysubgraph, and lastly we carefully classify the face identified by thehistogram of the subgraph information of the wavelet low frequency andhigh frequency. Through the experiment we find the measure is effectivewhen the face is standard, but we must improve on it in order to obtainan effect when the face is not standard.