Fingerprint image segmentation heavily influences the performance of every automated fingerprint identification system. Most of the feature extraction algorithms extract a lot of false features when applied to the noisy background area. Therefore, the main goal of these gmentation algorithm is to discard the background and reduce the number of false features. In this paper, we introduce a new approach to fingerprint segmentation based on the directional information. In conjunction with the size of nonoverlapping blocks, the directional variance of pixel number chosen in each block is used to segment fingerprint images. First, the direction image is estimated with a new orientation estimation algorithm. Then, in each block,enough pixels are chosen and the direction variance of pixel number is calculated. Finally, an adaptive threshold will be applied to determine the blocks of interest. This threshold is related with both the size and the pixel number of each block. Experiments show that the proposed method and manual segmentation perform equally well in rejecting false fingerprint features from the noisy background.This method is tested on the FVC2000 database and the experimental results have proven its efficiency.
展开▼
