基于高辨识度尺度不变特征的人耳识别方法  

Ear Identification Method Based on High Discriminative Scale-Invariant Features

作　　者：冯明[1] 

机构地区：[1]雅安职业技术学院,四川雅安625000

出　　处：《西南师范大学学报（自然科学版）》2015年第5期61-66,共6页Journal of Southwest China Normal University(Natural Science Edition)

基　　金：四川省高等教育人才培养质量和教学改革项目(川教函[2014]156号)

摘　　要：提出一种基于对人耳图像仿射不变特征进行小波变换的人耳识别方法.首先,提取人耳图像SIFT(尺度不变特征转换)特征,然后对SIFT特征向量进行一维小波变换获得更具辨识度的特征,最后利用余弦距离进行特征点匹配进而完成识别.实验结果表明,与一些传统的方法相比,本文提出的方法表现出更好的性能.Efficient feature extraction is the key to improve the accuracy of recognition .The scale-invari-ant feature transform （SIFT ） algorithm owns good invariance in the situation of affine transform ,noise and a certain degree light strength changing ;wavelet transform provides a sparse representation of a sig-nal .It has been shown that it closely matches with the human perceptual system .In this paper ,an ear i-dentification method based on wavelet transform of the Scale -invariant feature transform features has been introduced .First ,extract the SIFT features of human ear image .In addition ,one - dimensional wavelet transform is applied to the SIFT feature vector to obtain more discriminative feature .At last ,the cosine distance classifier is used for feature points matching and then complete recognition .Experimental results confirm that the high performance of the proposed method compared to some existing conventional methods .

关 键 词：人耳识别 尺度不变特征转换 小波变换 余弦距离 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]