摘要
在带有角度的人耳图像上提取有效特征一直是人耳识别的难点.本文提出一种基于Haar小波变换和规范型纹理描述子的人耳识别方法,即先对人耳图像进行Haar小波变换,然后利用更加合理的规范型纹理描述子,同时结合分块与多分辨率思想,共同描述经Haar小波变换后人耳子图像的纹理特征,最后用最近邻分类器进行分类识别.实验结果表明,Haar小波变换可以有效增强图像纹理基元的有效信息;利用规范型纹理描述子提取特征不仅速度快,而且具有很强的鲁棒性,尤其与分块、多分辨率方法相结合时,效果更为显著,明显优于经典的PCA和KPCA方法.
It has been an intractable problem for ear recognition to extract effective features from the posed ear images.This paper proposes a novel method based on Haar wavelet transform and uniform texture descriptors.Firstly,ear images are decomposed by Haar wavelet transform.Then using more rational uniform texture descriptors and combining simultaneously with block-based method and multi-resolution method describe together the texture features of ear sub-images transformed by Haar wavelet.Finally,the texture features are classified by the nearest neighbor method.Experimental results show that Haar wavelet transform can boost effectively up intensity information of texture unit.It is not only fast but also robust to use uniform texture descriptors to extract texture features.The recognition rate outperforms remarkably than that of the classic PCA or KPCA,especially when thinking block-based method and multi-resolution method into account.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2010年第1期239-243,共5页
Acta Electronica Sinica
关键词
HAAR小波变换
规范型纹理描述子
人耳识别
分块
多分辨率
Haar wavelet transform uniform texture descriptors ear recognition block-based multi-resolution
