基于相对熵和复小波变换的纹理图像检索

被引:5
作者
尚赵伟
唐远炎
房斌
郎方年
机构
[1] 重庆大学计算机学院
基金
高等学校博士学科点专项科研基金;
关键词
小波; 复小波; 纹理特征提取; 纹理检索;
D O I
暂无
中图分类号
TP391.41 [];
学科分类号
080203 ;
摘要
为了进一步提高纹理图像的检索性能,提出了一种基于复小波理论的特征提取算法,该算法根据复小波分解的特点,从小波系数角度出发,以每个高频子带复小波系数的实部和虚部系数直方图分布特性作为纹理特征,利用最大似然估计规则将特征提取和相似计算结合起来,以相对熵为依据进行相似计算。与单小波方法比较,该算法具有时移不变性、方向性信息多等特点。理论分析和在纹理图像检索的对比实验数据说明了复小波在纹理特征提取方面的性能优于单小波,且Kingsbury方法的性能优于Fernandes的方法。
引用
收藏
页码:541 / 548
页数:8
相关论文
共 10 条
[1]  
Wavelet-based texture retrieval using Generalized gaussian density and kullback-leibler distance. Minh N Do,Martin Vetterli. IEEE Transactions on Image Processing . 2002
[2]  
Texture analysis and classification with tree-structure wavelet transform. Chang T,Kuo C C J. IEEE Transactions on Image Processing . 1993
[3]  
A theory for multiresolution signaldecomposition:the wavelet representation. MALLAT S. IEEETrans Pattern Anal Machine Intell . 1989
[4]  
Three-dimensional attributesfor seismic interpretation[C/OL]. VAN SPAENDONCK R L C,HINDRIKS F M,FERNANDES F C A,et al. Annual MeetingAbstracts,Society of Exploration Geophysicists . 2000
[5]  
Texture classification using wavelettransform. GANESAN S A L. Pattern Recognition . 2003
[6]  
Texture features for browsing and retrieval of image data. Manjunath B S,Ma W Y. IEEE Transactions on Pattern Analysis and Machine Intelligence . 1996
[7]  
Image processing with complexwavelet[C/OL]. KINGSBURY N G. Phil Trans Royal Society London A . 1999
[8]  
Shiftablemulti-scale transforms. SIMONCELLI E P,FREEMAN W T,ADELSON EH,et al. IEEE Transactions on Information Theory . 1992
[9]  
Directionalscale analysis for seismic interpretation[C/OL]. VAN SPAENDONCK R,BARANIUK R G. Annual Meeting Abstracts,Society of ExplorationGeophysicists . 1999
[10]  
A new framework for complex wavelet transforms. FERNANDES F C A,VAN SPAENDONCK R L C,BURRUS C S. IEEE Transactions on Signal Processing . 2003