| Information security attracts more and more attention as the information technology develops. The traditional technique for security such as keys, passwords have failed to fulfill peoples' higher demand for protecting information, data and property, because of its easily lost and stolen. Biometrics, in particular with fingerprint, which have the characteristics of distinctiveness and permanence, can substitute keys, user names, passwords and other means to authenticate identity permanently. Meanwhile, they solve problems such as theft, loss and other issues, and provide a better performance of accuracy and security than traditional tokens. As the most mature biometrics identification technology, automatic fingerprint identification system has a variety of applications in forensic, information security, finance, electronic commerce and other fields. Fingerprint identification is now becoming an interdisciplinary area among pattern recognition, digital image processing and bio-information.This paper delivers studies on fingerprint sensor technology, pre-processing of fingerprint image, feature extraction and matching. R focuses on developing image acquisition and reconstruction for sweep sensor, proposing an algorithm for image segmentation based on fingerprint image direction map and block gray-scale information, improving Gabor filter for image enhancement, combining improved Poincare index and k-means in minutiae detection. At last, it develops a hardware module of automatic fingerprint identification system for embedded application. |
PDF Full Text Request