Font Size: a A A

Automated Fingerprint Identification Technology Research

Posted on:2011-06-11Degree:MasterType:Thesis
Country:ChinaCandidate:J K WangFull Text:PDF
GTID:2208360305993686Subject:Circuits and Systems
Abstract/Summary:PDF Full Text Request
Fingerprint recognition is one of the technologies of biometric authentication. Compared to traditional personal identity authentication, it has the advantage of safety, accurateness and convenience and is widely applied in practice. But current fingerprint recognition algorithms still show many shortages at speed, accurateness, FAR, FRR, robustness and so on. The thesis has studied several key technologies of fingerprint recognition, mainly including fingerprint segmentation, fingerprint enhancement, singular points extraction, and fingerprint feature matching and so on. Moreover, an automatic fingerprint recognition system platform for development is achieved. The thesis mainly makes the following research results:1. In the aspect of fingerprint image preprocessing, firstly, two stages segmentation strategy is proposed to segment fingerprint. At first blanks are taken away, and then the separation for the remained clear texture area and noise area is taken as the problem of two value classification and can be classified by using simple linear classifier. Secondly, fingerprint enhancement by Hermite filter is studied. The algorithm calculates ridge frequency by using the frequency spectrum characteristic of fingerprint image block, and uses the good band-pass property to enhance the fingerprint texture definition. To better avoid blocking effect in singular point area, the algorithm sets changeable angle bandwidth.2. In the aspect of fingerprint feature extraction, firstly, a hierarchical extraction method is used to extract singular points. At first the location of singular point area is found coarsely by using fingerprint curvature, and then the location and orientation of singular point are accurately extracted by using complex filter technology at point orientation field in singular point area. Secondly, a knowledge-based method is used to remove fingerprint false features. At first every kind of false feature structure is defined, and different methods are specialized to deal with the corresponding structures. At last, different kinds of false features are removed by using corresponding method in reasonable order.3. In the aspect of fingerprint feature matching, the fingerprint minu- tiae matching algorithm based on the local and global structures is improved. The thesis analyses the shortage of original fingerprint matching algorithm:one is that local structure includes unstable feature; and another is that fixed bounding box is used. To overcome these shortages, the thesis uses minutiae sort-index instead of minutiae type, which can improve the stability of local structure; and use bounding box with changeable size for global matching, which is better suited to nonlinear deformation.4. An automatic fingerprint recognition system platform for development is realized by programming in VC6.0 environment. The platform is designed by using object-oriented thinking, and integrates many kinds of algorithms for fingerprint recognition system, which can be used to study the algorithms of automatic fingerprint recognition.
Keywords/Search Tags:fingerprint recognition, fingerprint image preprocessing, fingerprint feature extraction, fingerprint feature matching
PDF Full Text Request
Related items