Algorithm Based on Point Feature for Fingerprint Image Segmentation

The segmentation of fingerprint images is very important to automatic fingerprint recognition system. To decide which part of the image belongs to the foreground and which part to the background, different methods have been proposed. Some of the traditional methods are introduced and implemented in this paper and a novel pixel based fingerprint segmentation algorithm is proposed. Two kinds pixel features, being the coherence of direction and the variance of gray level are used in the segmentation and fuzzy c-means clustering algorithm is used to select threshold. Morphology technology is applied as post processing to obtain a smooth contour line. The experimental results demonstrate the effectiveness of the proposed method, especially in low quality images, in terms of including less background and excluding less foreground. In addition, this robust segmentation algorithm is capable of filtering efficiently spurious boundary.