Image Enhancement via Multiple Canonical Correlation Analysis

被引:0
作者
Polat, Ozgur Murat [1 ]
Ozkazanc, Yakup [2 ]
机构
[1] ASELSAN, Mikroelekt Gudum & Elektroopt Grubu, TR-06011 Ankara, Turkey
[2] Hacettepe Univ, Elektrik & Elektronik Muhendisligi Bolumu, TR-06800 Ankara, Turkey
来源
2013 21ST SIGNAL PROCESSING AND COMMUNICATIONS APPLICATIONS CONFERENCE (SIU) | 2013年
关键词
Canonical Correlation Analysis; Canonical Variates; Scene Understanding; Image Enhancement; SETS;
D O I
暂无
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
For image understanding and performing detection, recognition and identification functions, different features representing the scene should be extracted from different images of the scene. From the images of the scene captured by different sensors, such as operating at different bands, different features can be obtained. For sensor fusion, first the difference in the information content of these separate data should be assessed. In this study, demonstration of the use of Multiple Canonical Correlation Analysis (MCCA) for information extraction from the multi-sensor data is provided. From the registered data captured with three different cameras, multiple images are obtained by pixel shifting methodology and analyzed via MCCA. The scene details are obtained from the canonical variates and the level of mutual information of these new data sets is determined via canonical correlations.
引用
收藏
页数:4
相关论文
共 7 条
[1]  
Anderson T. W., 1984, An introduction to multivariate statistical analysis, V2nd
[2]  
[Anonymous], 1998, THESIS LINKOPING U
[3]  
[Anonymous], 1973, Pattern Classification and Scene Analysis
[4]   Relations between two sets of variates [J].
Hotelling, H .
BIOMETRIKA, 1936, 28 :321-377
[5]  
KETTENRING JR, 1971, BIOMETRIKA, V58, P433, DOI 10.1093/biomet/58.3.433
[6]  
Lewis J. J., 2006, P 9 INT C INF FUS FU
[7]  
Via J., 2005, 13 EUR SIGN PROC C E