关键词: 生物特征识别, 掌纹识别, 感兴趣区域, 坐标系

Abstract: Palmprint recognition，a kind of biometrics，recently receives many researchers’ attention because of its low resolution，cheap devices and easily accepted by the users.The key step that finding a region of interesting（ROI） and segmenting from the background of the palmprint image is important for the feature extraction and further matching.This paper presents a new method of extracting the ROI.Firstly，a global threshold is used to binarize the original palmprint image，and the edge of the palm is extracted.Then，a new coordinate system is settled according to the slope making by the two gaps between the first and middle fingers，the ring and little fingers，and the centre of the palm.Finally，a fixed 128×128 subimage is extracted as ROI.Compared with other methods which correlative papers presented，the experimental results demonstrate the effectiveness and accuracy of the proposed method.

Key words: biometrics, palmprint recognition, region of interest（ROI）, coordinate system