Research In Pre-processing Of Fingerprint

The technology of fingerprint recognition is used widely in biometric recognition technology. Sometimes the result of recognition isn't satisfied, so researching in algorithm of fingerprint recognition has far-reaching signification. This paper mainly stress on the algorithm of fingerprint pre-processing.In this paper, a lot of learning literature about fingerprint recognition technology recently is analyzed firstly and the characteristics of fingerprint are construed. Some newly reasonable methods are put forward that based on the characteristic of fingerprint image, which is picked up with capacitive sensor. The methods include fingerprint segmentation, fingerprint binarization, fingerprint thinning, and feature extraction etc. In this paper, some reasonable pre-processing methods based on the characteristic of fingerprint image are put forward, and processed results are credible. Firstly, some kinds of segmentation are discussed, and the algorithm of segmentation based on directional information is chosen, the method is combined with square difference and directional information. And segmentation result can be received well. Secondly, after comparing some methods of binarization, improved binarization algorithm is chosen. Analyzing the result of improved algorithm and dynamic binarization, it concludes that improved algorithm is more feasible. Thirdly, after comparing virtue and disadvantage of the methods of OPTA and Sherman, Sherman thinning is chosen. The results prove that Sherman is more feasible. Lastly, it is proposed that the algorithm of feature extraction based on connective concept. And it is introduced that the concept of matching.In this paper, the methods of pre-processing have good results and run fast and easily to be programmed, which have effectively solved the problem identification of fingerprint.