Research Of Fingerprint Recognition Algorithm Based On Ridge Vectorization

Nowadays, fingerprint (FP) identification is becoming one of the most attractive, convenient and reliable technology of identification. Traditional fingerprint identification algorithms based on minutiae only use a small part of image information, that is the position and orientation of the minutiae, as character to match, which lost other abundant information contained in FP image. Also, identification of the two FPs which are in different sizes, different angles and positions, some solutions will unify their size, angle and position first. It makes the algorithm very complicated and lowers the accuracy of the identification.On the basis of former research, an automatic FP identification system is realized, which is made up of three parts including the FP image preprocess, the image vectorization and the fingerprint matching based on ridge curvature. Firstly, the FP image is preprocessed using a few classical image process algorithms including image model, image normalizing, image filter, binarization, thinning and post-process of the thinned FP image. Secondly, a vectorization technology is introduced into FP images in order to describe the FP ridges exactly. Lastly, a new FP matching algorithm based on the immutability of the curvature of the points on FP ridges is approached.The results demonstrate that the preprocess algorithms remarkably reduce the noise interference and eliminate the fake characters by the post-process of the thinned FP image. The vectorization of the FP image not only describes the FP ridges exactly, makes the FP image undistorted when zoomed in, but also reduces the storage space of the FP image. The final matching results based on the ridge curvature indicate that this algorithm has a very good achievement with different FP images and is invariant to rotation.