基于沟回结构特征的三维人耳识别方法  

3D human ear recognition method based on auricle structural feature

作　　者：王凯[1] 穆志纯[1] 

机构地区：[1]北京科技大学自动化学院,北京100083

出　　处：《仪器仪表学报》2014年第2期313-319,共7页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(60973064);高等学校博士学科点专项科研基金(20100006110014);北京市自然科学基金(4102039);北京市教育委员会重点学科共建项目(XK100080537);中央高校基本科研业务费(FRF-SD-12-017A)资助项目

摘　　要：人耳的生理结构含有丰富的身份鉴别信息,但由于其提取困难,现有的基于局部曲面特征的三维人耳识别并未对此进行利用。提出了一种新的三维人耳表征方法,称之为沟回结构特征。利用曲面变化量度量曲面的凹凸特性,进而提取出人耳沟回的关键结构信息。基于提取的沟回结构特征利用迭代最近点ICP算法进行三维人耳的粗对准,并进一步在三维人耳点云上进行ICP精对准。在UND生物识别图像库集合J2上的实验获得了较现有基于曲面表征和ICP算法的三维人耳识别方法更高的识别率和更短的耗时。Physiological structure of the human ear contains rich information for personal identification. However, no existing local surface representation based 3 D human ear recognition methods take the structure into account, mainly because of the difficulty in its extraction. In this paper, a novel 3 D human ear representation called 3 D auricle struc- tural feature （3DASF） is proposed. The Surface Variation is used to measure the feature of the surface ,and then the auricle structural information of the ear is extracted to obtain the 3 DASF, which is used to coarsely align the probe and gallery ear pairs with iterative closest point （ICP） algorithm. Subsequently, the ICP algorithm is performed to a- chieve a refined alignment. The experiment was conducted on the university of notre dame （UND） collection J2 data- set,and the experiment results show that the proposed method outperforms the state-of-the-art 3D ear recognition methods that use local surface representation and ICP in terms of recognition rate and time consumption.

关 键 词：三维人耳识别 三维特征提取 沟回结构 迭代最近点 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]