Research On Methodology Of Fuzzy Hand Feature Extraction

In today's information times, the traditional identification methods can be easily lost and fake, which can't satisfy the demand of people's life the contemporary social. Accurately identify a person's identity and protect the security of information have become key social problems which must be solved. The biometric technology is the best solution, which is more convenient and safe, accurate then traditional identification. Hand shape recognition technology has many advantages, such as low cost, simple operation, and quick authentication, etc, which is widely used in the entrance guard, checking attendance system.This paper takes research in hand forms images with fingers stretching out under the non-contact with opened naturally condition. First, establish clear gallery and different fuzzy degrees of hand images. Then it puts forward automatic location methods of feature points based on hand contour, at the same time gives shape characteristic parameters calculation method, and then through experiment results analyze the positioning precision. Standard deviations focus on 2, which indicate that the positioning of high stability. The causes of blurred image galleries are analyzed, and then fuzzy hand form is established and the method of feature point positioning is put in this paper. Exact the fingers'lengths, explore fingers relative length's change cased by distance. Calculate standard deviation and maximum error using the data of relative lengths of fingers, and then analyze the stability of positioning based on fuzzy image. The result displays relative lengths of standard deviations focus on 1%. Experiment shows that this position method can exact effectively feature. At the last, different characteristic vectors were exacted aiming to clear gallery, different characteristic vectors selected to carry out recognition experiment, using the matching feature vector realize hand form authentication. According to the fuzzy hand fractal image, select registered effective sample to obtain higher recognition rate of fuzzy hand fractal image.Typical biometrics system contains registration module, feature extraction module and recognition module, which are designed and combined with each other through the interface to realize fractal independent hand form recognition system. Finally successfully design the man-machine interface, and complete shape recognition/hand real-time authentication system.