FEATURE EXTRACTION OF FINGERPRINT IMAGES

Abstract

As the demand of information security increase, the application of fingerprint has been widely used. The application of fingerprint included entrance monitoring system, identification of user status before entering the computer system and so on. In fingerprint verification system of police department, the fingerprint of crime that is hard to recognize by naked eyes can also be verified, but, it will spend a large time to analyze. Fast and exact recognition is important in on-line fingerprint verification system. If the quality of image is poor after examination, we will enhance it in order to have a better one. The main purpose of image enhancement is to improve low quality images, so that it can avoid the users to re-provide their fingerprint just because it is no good impressed. Therefore, if the image is hard to recognize after image enhancement, i.e. the quality of image is poor, thus, it will be rejected to recognize. We have defined a simple criterion area to examine the captured fingerprint images. Moreover, in order to reduce the computational flow and time, we do preprocessing in the criterion area instead of the whole fingerprint image. In image preprocessing, generally, Gray-Level is used to reduce computational flow. However, we suggest another approach as follow. After we examine the captured fingerprint images, we can calculate the value of the minimum red-value subtract the maximum blue-value. It can almost help us to capture the “ridge” of the fingerprint image. In pruning, we propose a simple concept － slope to trim the image.