Lines are the most essential and discriminative features of palmprint images,which motivate researches to propose various line direction based methods forpalmprint recognition. Conventional methods usually capture the only one of themost dominant direction of palmprint images. However, a number of points inpalmprint images have double or even more than two dominant directions becauseof a plenty of crossing lines of palmprint images. In this paper, we propose alocal multiple directional pattern (LMDP) to effectively characterize themultiple direction features of palmprint images. LMDP can not only exactlydenote the number and positions of dominant directions but also effectivelyreflect the confidence of each dominant direction. Then, a simple and effectivecoding scheme is designed to represent the LMDP and a block-wise LMDPdescriptor is used as the feature space of palmprint images in palmprintrecognition. Extensive experimental results demonstrate the superiority of theLMDP over the conventional powerful descriptors and the state-of-the-artdirection based methods in palmprint recognition.
展开▼
