Abstract: Minutiae extraction is a critical problem in the research of automatic fingerprint recognition Though there are many methods based on the analysis of a skeleton image or original gray image, the problem has not been solved perfectly In this paper, a method is presented which extracts minutiae from the binary image Image segments are first extracted from the binary image by run length code matching Then the minutiae are detected through the analysis of the segments’ structure and their inter relations The direction of minutiae is determined by the line testing method Experimental results show that compared with the method based on the skeleton image, the proposed method is more efficient and can improve the accuracy of fingerprint matching