Improved adaptive threshold algorithm for fingerprint segmentation based on multiple features

Segmentation is an important step in a fingerprint recognition system. Different methods have been proposed for fingerprint segmentation. In this paper, some of the most important segmentation methods are introduced and an improved adaptive threshold algorithm based on multiple features is introduced and implemented. Morphology technology is applied as post processing to obtain a smooth contour line. The FVC2004 is selected as a standard database, to implement each method. Experimental results prove that the algorithm provides very accurate segmentation even for low quality images.