Automated Fingerprint Identification System

An accurate automatic personal identification is critical to a wide range of application domains such as access control, electronic commerce, and welfare benefits disbursement. Traditional personal identification methods suffer from a number of drawbacks and are unable to satisfy the security requirements of our highly inter-connected information society. Biometrics refers to automatic identification of an individual based on his/her physiological and/or behavioral traits. While biometrics is not an identification panacea, it is beginning to provide very powerful tools for the problems requiring positive identification.Automatic fingerprint identification is one of the most reliable biometric technology. This dissertation focuses on Automated Fingerprint Identification Systems(AFIS), Based on current theoretic framework, a complete AFIS has been accomplished. Attention has been given to some key problems such as fingerprint image pre-processing, features extraction and minutia matching for further improvement. The main contribution of this thesis includes:Two methods of segment fingerprint from background are studied: Variance method and direction method. In addition, we have been improved the variance method.A fingerprint enhancement algorithm based on gray-level image is introduced, in this algorithm, good band-pass performance of Gabor filter is employed, ridge direction and ridge frequency of block fingerprint image is used as parameter, Gabor filter is used to filter and enhance the image quality.Three methods of binary image have been realized: Global threshold method, local threshold method and direction method. Through comparing experimental result, we find a kind of method most suitable for this system.Several methods of thinning are analyzed: Traditional thinning algorithm, serial and parallel thinning algorithm and improved OPTA thinning algorithm. We bring out a new composite thinning method. Experimental results show it iseffective.An improved algorithm for features extraction is brought out. After a set of original features is directly extracted from thinned fingerprint images, various noises in fingerprint images and their properties are analyzed, special algorithm are designed with respect to various noises to delete pscudo features. The remainders of original features are viewed as intrinsic features.A fingerprint matching algorithm called local structure and global structure approach is proposed. We also introduce the ridge information in the global matching stage , Experimental results show it improved the fingerprint performance effectively.In the last part of this dissertation, conclusion is made and the direction of future research is further proposed. With the development of science and technology, fingerprint recognition system will bring more facility for people daily life in the nearest future.