基于CCA的图像语义特征提取的分析与研究

被引:3
作者
韩昌刚 [1 ]
郭玉堂 [2 ]
机构
[1] 安徽大学计算机科学与技术学院
[2] 合肥师范学院计算机科学与技术系
基金
安徽省自然科学基金;
关键词
图像语义; 典型相关分析; 局部二值模式; 特征参数; 特征融合;
D O I
暂无
中图分类号
TP391.41 [];
学科分类号
080203 ;
摘要
为了提高图像语义特征提取的精确度,克服目前大部分图像语义特征提取算法中,因图像特征提取不当,导致特征参数不能全面反映图像语义的问题,提出了一种基于典型相关分析(CCA)的特征融合的图像语义特征提取方法。该方法首先采用圆形对称邻域取代传统的矩形邻域的方法,对局部二值模式(LBP)纹理特征进行了改进,然后采用高维小样本下典型相关分析对可伸缩颜色描述算子的颜色特征和改进的LBP纹理特征进行特征融合。实验结果表明,所提出的方法明显提高了图像语义特征提取的精确度,能有效地建立图像的低层特征与语义特征间的一致性。
引用
收藏
页码:1938 / 1942
页数:5
相关论文
共 13 条
[1]  
A survey of con-tent-based image retrieval with high level semantics. Liu Ying,Zhang Dengsheng,LUGuojun,et al. Pattern Recognition . 2007
[2]  
多元统计分析导论[M]. 人民邮电出版社 , (美) 安德森 (Anderson, 2010
[3]  
Architecture and analysisof color structure and scalable color descriptor for real-time video inde-xing and retrieval. CHANG Jing-ying,LIAN C J,CHEN L G. Proc of the 5th Pacific Rim Conference onAdvances in Multimedia Information Processing-Volume PartⅡ . 2004
[4]  
Performance ofMPEG-7 edge histogram descriptor in face recognition using principalcomponent analysis. RAHMAN S,NAIM S M,ALFAROOQ A,et al. Proc of the 13th International Conference onComputer and Information Technology . 2010
[5]  
Medical image retrieval using local bi-nary patterns with image Euclidean distance. XU Xian-chuan,,ZHANG Qi. Proc of Internation-al Conference on Information Engineering and Computer Science . 2009
[6]  
Integral localbinary patterns:a novel approach suitable for texture-based object de-tection tasks. PEREIRA E T,GOMES H M,DeCARVALHO J M. Proc of the 23rd SIBGRAPI Conference on Graph-ics,Patterns and Images . 2010
[7]  
Local two-dimensional canonical correlation analysis. HAI Xian-wang. Signal Processing . 2010
[8]  
Slim-tree and BitMatrix in-dex structures in image retrieval system using MPEG-7 descriptors. AEAR E,ARSLAN S,YAZICI A,et al. Proc of International Workshop on Content-Based MultimediaIndexing . 2008
[9]  
An improved edge detection innoisy image using fuzzy enhancement. CHEN Xiang-tao,CHEN Yu-juan. Proc of International Con-ference on Biomedical Engineering and Computer Science . 2010
[10]  
Learning object detection from a small number ofexamples the importance of good features. LEVI K,WEISS Y. Proc of IEEE Comput-er Society Conference on Computer Vision and Pattern Recognition . 2004