Research On Fingerprint Recognition Algorithms

Automatic identification technique has been an emerging approach in real life. Due to its strong reliability, biometric authentication has drawn much attention ,a large number of scientific researchers and technology enterprises were engaged in researching and developing related products.As one kind of key biometrics technology, although, fingerprint recognition has been extensively studied and many advances have been made on it, there are still many problems expected to be solved which are shown in actual applications and evaluations. As a consequence, many academies and industries have been making an in-depth research on fingerprint recognition technologies. This thesis has studied deeply several key issues in fingerprint recognition technology and bring forward new arithmetic ideas, realization methods and experiment results mainly includes the following contributions:This paper has developed an in-depth research of some key issues in automated fingerprint identification technology and has presented the improved algorithm and simulation results. The main findings consist of the following: In view of problems that the noisy region's direction information of the fingerprint is difficulty to extract accurately we executed robust algorithm to adjust the direction image after the gradient method which can adjust larger fingerprint fuzzy regions. In fingerprint image segmentation, the analysis of variance and direction of the fingerprint has been made and a smooth contour fingerprint segmentation algorithm has been designed, Simulation results show that the algorithm can accurately extract fingerprint images from the background, and can remove the edge of non-normal interrupt endpoint; In the respect of thinning in the fingerprint, The principle of image refinement is studied and the paper displays a barycenter based thinning methods, with which, the fingerprint image thinning algorithm ensure that a single fingerprint ridge pixels, connectivity, no spur, maintaining the refinement for non-linetype image at the same time; Finally, for the problem on the difficulty in determination of fingerprint alignment reference point in the point pattern matching, affine transformation of the reference point congruent triangles has been used to overcome the elastic deformation in fingerprint images to some extent.