Fingerprint Image Preprocessing Algorithm Research

The technology of fingerprint recognition is an active subject in the area of pattern recognition, it has broad application. The fingerprint image preprocessing is the first step of fingerprint identification system, the quality of the fingerprint image preprocessing affects directly the result of identification system, and the core of the whole system is images enhancement, binarization and thinning. This thesis mainly focuses on the algorithms of enhacement, binarization and thinning, while in-depth research is carried out into the algorithms of directional image which will be used in the process of enhancement, binarization and thinning. By summarizing and digesting the researching findings by scholars from both home and abroad, this thesis realized many kinds of directional image, enhancement, binarization and thinning algorithms by programming. At the same time, in the thesis some algorithms are improved in some aspects so as to obtain a better quality of operation.The improved algorithms including:(1) Fingerprint image enhancement: At first,this thesis introduce some algorithms about the orientation field algorithm which is used in the next procedures, and realize the orientation field based on linear structure tensor. After the analysis and realization of two commom method of fingerprint image enhacement, this thesis take the method of PDEs in the fingerprint image enhacement, improve the algorithm of J.Weickert's nonlinear diffusion PDE model, rebuild the diffusion tensor , and make the diffusion in the complex area much better.(2) Fingerprint image binarization: We did some research on the algorithm on image bianrization, and advanced a new method of image binarization based on orientation field and frequency information. As the experiment result , The new algorithm outperforms the other method in noising image.(3) Fingerprint thinning: This thesis improve the algorithms on fast thinning method and Improved OPTA thinning method, solve the problem of incomplete thinning in the fast thinning method, and also solve the problem of incomplete thinning and thinning not smoothly enough in the Improved OPTA thinning method.