Segmentation of Fingerprint Images Using Minimal Graph Cuts

Segmentation of fingerprint image is to extract the region of interest (ROI) of image and highly influences the performance of Automatic Fingerprint Identification System (AFIS). For each image block, either background or foreground label should be determined. In traditional methods, the label of an image block is only based on the features from this block itself such as local gray variance and local orientation coherence without considering the effect of its neighbors' labels. In this paper, we present an efficient technique for fingerprint image segmentation using minimal graph cuts with considering the effect of the neighbors' labels. The advantage of the proposed method is to make the labeling "smoother" and "more coherent" by minimizing the amount of foreground/background boundary. Experimental results demonstrate the good performance of the proposed method.