掌纹识别 定位分割 形态学腐蚀膨胀算法 matlab源码（Palmprint recognition, location segmentation, morphology, corrosion, expansion algorithm, matlab source code）.docVIP

掌纹识别 定位分割 形态学腐蚀膨胀算法 matlab源码（Palmprint recognition, location segmentation, morphology, corrosion, expansion algorithm, matlab source code） The positioning of palmprint recognition morphological segmentation of corrosion expansion algorithm matlab source temperature Changzhi Southwest Jiao Tong University on Thursday 02 October 2008 14:09%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Morphological corrosion dilation algorithm for preprocessing procedures of% Function: palmprint images % Author: Wen Changzhi % Date: 2006/12/7 Sichuan Key Laboratory of signal and information processing, Southwest Jiao Tong University, Copyright: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % % % of the second algorithms employ corrosion and dilation operations in morphology. First of all, the image is low-pass filtered to eliminate the interference of small noise. When filtering, the traditional median filter is not used, but the sequential statistic filter ordfilt2 is adopted. % imm=ordfilt2 (IM, N, ones (N, M)); % upper expression indicates that the values of pixels in the range of ones (N, M) are sorted in sequence, and the values of the pixels at the N point are used instead of the pixel values of the points, The same as the median filter, the filter does not blur the edges of the image, and preserves the edges of the image while filtering the noise. After setting the threshold value, the palmprint image is transformed into a two valued image. The threshold should be 15 (or 20) from experience. % of the two value of the image expansion operations, and then corrosion operations, the results will be subtracted from the original image, the remaining points on the image is the corner. % first dilation operations and then corrosion operations, and SE represents the structural elements used in expansion and etching. The determination of% SE is the key to this corner detection. The shape of the SE should be similar to the shape of