Research On Low-quality Fingerprint Recognition

Automated Fingerprint Identification Technology is currently the most widely used as a biometric identification technology, Which broadly include the fingerprint image pre-processing (segmentation fingerprint, fingerprint enhancement, fingerprint refinement), also include feature extraction and matching three stages. So far, a lot of literature has been conducted in-depth studies of it. In the actual process of fingerprints collecting and recording, because of the changes of compression, the collector's own shortcomings, such as human factors, In order to ensure the quality of the robustness of the Fingerprint, We need to in-depth study of the improvement of the fingerprint identification.In order to realize the fingerprints enhancement in my Graduation Project, This paper has proposed two algorithms, One algorithm is based on a Log-Gabor filter, The method works well and solves the problems such as empty, broken. Another fingerprint image enhancement method is based on the direction filter. The key of this method is the realization of the filter design, as well as the realization of filter rotation. The processing speed of this method is very fast,at the same time the ability to remove the noise and repair the empty noise, fracture is also very good. This design use VC++ programming to achieve the above algorithms, include another two refining algorithms,and compare the treatment effect.This paper also proposed a new method of fingerprint identification base on Pulse Coupled Neural Network this method has a stronger ability of segmentation for different types of images as well as reduces the number of iterations.