Modified PCNN Filtering for Fingerprint Enhancement

Enhancement is an important step in fingerprint system. This paper presents one fingerprint enhancement algorithm, which is based on a modified pulse coupled neural network (PCNN). It preprocesses fingerprint to obtain image and orientation masks, and then it iteratively enhances the image by PCNN filtering by the constraints of these two masks. Experiments on FVC2004 databases show that this algorithm is able to enhance fingerprint effectively, and it could also significantly improve the recognition performance of a fingerprint system.