Research On Face Recognition Based On Skin Color And Artificial Neural Network

Biometrics is a reliable and stable identification, which relies on unique human physical or behavioral characteristics. Among them, the face recognition biometrics technology is also a recognition of the most dominant technology. Consistent with the human habit of face recognition, it has a natural friendly, directly and effective features, easy to be accepted. Specifically the first input image to determine whether there is face, and then operate to determine if there is face to face to determine the information, the detection and identification.With the rapid development of science and technology, Face Recognition with rapid development, while face recognition has also been a major breakthrough. For greater investment in research on face recognition, there has been a variety of face recognition system, each research also made a number of effective face recognition.Face Recognition process generally includes face detection, feature extraction, and ultimately identify the three processes. In order to make this a more effective detection and recognition of face images, the image is preprocessed. Then, the processed image using skin color detection method to detect human face, presented a color image processing and detection methods and theories related to knowledge. Finally, the image recognition is proposed QGA-BP neural network method. The main contents are as follows:1.Overview of the recognition of meaning and value of the background, status and application of face recognition, the main problem facing so on.2.Introduced the color image processing common color space and color model, and face detection and recognition of the various methods are analyzed and compared.3.This article focuses on the recognition process:First face image preprocessing operation;Then YUV-KL color space, face detection based on skin color model; Combined with prior knowledge of facial features, employing some of the outstanding features of the face to people who have been detected person face region confirmation, and determine the person face region finally; KL transformation method based on extraction of facial features, for the following to better the human face Ready identification; The final step is to recognize human faces, we use the QGA-BP neural network to recognize faces. One of the quantum genetic algorithm and BP neural network are introduced, focusing on the combination of the two constitutes an effective identification method. And only use BP neural network for face recognition and face recognition using a combination of both for a comparative analysis.4.Summarizes the research work of this paper,and the prospect of further work.