Study On Palmprint Recognition With Fuzzy Methods

Unique characteristic and relative stability make the biological characteristics of the individual not easily be forged or counterfeit. Therefore the use of biometric authentication technology are promising .Biometric palm print recognition is an emerging area which attracts a growing number of researchers and has become hot in the field of biometrics research.This paper summarized the research of palm print recognition technology, analyzed the advantages and limitations of the methods having been proposed, and studied the key technology and the core algorithm of the palm print identification system for image pre-processing (including the regional palm print image positioning, normalizing), palm print feature extraction, feature matching and so on. The results are as follows:Firstly, palm print images are from Polyu-Online-Palmprint Database in the Hong Kong Polytechnic University Biometric Research Center. Pre-processing was based on the contour feature point location methods, the experiments showed that this method can effectively position and partition, to some extent, reduce the unnecessary noise.Secondly, palm print recognition based on the fuzzy similarity was proposed, after analyzing the characteristic of palm print image, the membership function was proposed to establish the similar palm print image matrix, and then Euclidean distance was used to identify the image. Principle of fuzzy can effectively resolve the uncertainty of gray-scale images, the geometry and the knowledge.Finally, the whole algorithmic of recognition process was simulated using Matlab software, meanwhile, the recognition result, de-rotation and de-shift capability were tested. The experimental results showed that fuzzy similarity-method used in the palm print recognition, to some extent, can inhibit the translation and rotation on the recognition results and have better real-time capabilities and robustness.