ناحيه بندي خودكار اثر انگشت

Automatic Fingerprint Segmentation

Fast and accurate segmentation of fingerprints is essential to each automatic fingerprint identification / verfication system. In this research we have developed new approach to fingerprint restoration and segmentation using directional properties of ridge patterns. Based on Man filters for image segmentation and binerization, a few directional filters are designed for different directions. Using these filters as classifiers, each pixel of fingerprint image is assigned either to ridges or to background valleys. Furthermore for each ridge point an approximate direction is calculated and stored in segmented image. This directions are then smoothed in some windows around each pixel to produce the final restored and segmented image with additional ridge direction information, The accuracy and speed of this method is tested for a large number of fingerprint images with different initial qualities and is found to be excellent compared with some other methods. In addition, an extremely fast and accurate algorithm for thinning the fingerprint images is presented.