一种掌纹线特征的提取方法  

A Method for Extraction of Line Features of Palmprint

作　　者：宋炯[1] 林喜荣[1] 包桂秋[1] 申晶[1] 

机构地区：[1]清华大学精密仪器与机械学系,北京100084

出　　处：《计算机工程与应用》2004年第12期32-33,57,共3页Computer Engineering and Applications

摘　　要：文章尝试了一种对掌纹线特征,即屈肌线特征和大的皱纹特征进行提取的方法。该方法先将掌纹图像划分成若干小方块,然后通过计算像素点的方向数确定每个小方块的主方向,进而得到每个小方块的沿主方向的灰度分布曲线,并以此进行区域判别,保留纹线区域,屏蔽非纹线区域。从实验结果可以看出,掌纹的绝大部分线特征得以保留,同时抑制了大部分的噪声。可见这种方法是可行而有效的,为基于掌纹的身份自动鉴别技术的研究积累了有益的经验。This paper tries a method for extraction of line features,i.e.principal-line features and large wrinkle features,of palmprint.This method firstly divides palmprint image into several little blocks,then ascertains the main trend of ev-ery block by calculating the trend number of pixels,and then acquires the gray scale distribution curve along the main trend of every block,and makes area distinguish by it,retain s dermal area,shields none dermal area.The experiment re-sult shows that the great mass of line features of palmprint is retained while the mass of noise is restrained.So this method is feasible and effective,it accumulates beneficial experience for the study of automatic identity discrimination technique by palmprint.

关 键 词：掌纹 身份鉴别 方向数 主方向 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]