自动找点的掌纹图像定位分割方法  

Palmprint image positioning and segmentation methods based on auto finding points

作　　者：许爽[1,2] 索继东[1] 赵继印[2] 

机构地区：[1]大连海事大学信息科学技术学院,辽宁大连116026 [2]大连民族学院信息与通信工程学院,辽宁大连116600

出　　处：《计算机工程与设计》2011年第9期3102-3105,共4页Computer Engineering and Design

基　　金：中央高校基本科研业务费专项基金项目(DC10010103);辽宁省教育厅科学技术研究基金项目(L2010094)

摘　　要：重点研究掌纹图像在产生旋转和平移情况下的定位分割方法。对采集的掌纹图像做中值滤波后得到二值图像,跟踪二值图像边界,转换边界点重新建立只显示边界新图像,最后分析手掌形态,分离手指与手掌,利用去孤立点的自动找点方法提取食指与中指、无名指与小拇指两处定位点,并以此对图像做旋转校正,建立直角坐标系,确定图像的感兴趣区域。实验结果表明,该方法的定位准确率高,有利于进一步对掌纹图像进行特征提取和特征匹配。The positioning and segmentation of palmprint image is discussed in detail,which involves rotation and transition.Firstly,palmprint image used median filter is converted binary image.Then the boundary of binary image is tracked and a new image is recreated,which shows only the border.Finally,the fingers and palms are separated and the two positioning points of the index finger and middle finger,ring finger and little finger are found.Thus,the palmprint image is rotated and corrected based on the two positioning points and coordinate system is created according to positioning points,which determines the region of interest（ROI）.The experimental results show that positioning accuracy is higher and is beneficial to further feature extraction and feature matching.

关 键 词：掌纹图像 定位 分割 二值图像 感兴趣区域 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]