Research On Fingerprint Recognition Algorithm Based On Structure Feature Matching

Fingerprint identification is one of the most important and effective biometric technology which has been paid amount of attention. It is the significant part of individual authentication. Nowadays many fingerprint products have been offered, however, most of the key methods wouldn't be publicized because of commercial benefits. Besides, more requirements on the fingerprint identification system are made because of the society's development. So studies on the theory and application of fingerprint recognition have significant meanings both in theory and application.Base on plenty of papers, this dissertation makes some research on common fingerprint recognition,which including algorithms on fingerprint preprocessing, especiaolly on feature extraction and feature matching.The main points of the dissertation are as follows:1.In the process of computing fingerprint direction image,analyze neighbor direction mask method and min average covariation block direction estimation method.Focusing on the problem of point direction with low accuracy,propose a improved scheme in which applicate the min average covariation algorithm to compute the direction of every point.This scheme improves the accuracy of direction image.2.In the process of fingerprint feature extraction,study two methods on extracting fingerprint feature:preprocessing-based extraction and Gray fingerprint image-based direct feature extraction. The experimental results have been compared.3. In the process of fingerprint matching, a new fingerprint matching method is presented. This algorithm constructs the structure features based on the number of ridge lines and D-value of angles between minutiae and core point. The experimental results demonstrate that it is robust to rotation, excursion, distortion and condensation.4. Use an improved Bresenham draw line algorithm to compute the ridge number between two points. It can calculate the number of ridges between two points exactly because it could avoid missing or tell wrong ridges.All algorithms mentioned in this thesis are carried out with Visual C++ on the computer.The experiment results is satisfied.