Feature Fusion Using Three-Dimensional Canonical Correlation Analysis for Image Recognition
A new feature fusion method, namely three-dimensional canonical component analysis (TCCA), is proposed in this paper. It is an extension of traditional canonical correlation analysis (CCA) and two-dimensional canonical correlation analysis (2DCCA). The method can directly find the relations between two sets of three-dimensional data without reshaping the data into matrices or vectors, and dramatically reduces the computational complexity. To evaluate the algorithm, we are using Gabor wavelet to generate the three-dimensional data, and fusing them at the feature level by TCCA. Some experiments on ORL database and compared with other methods, the results show that the TCCA not only the computing complexity is lower, the recognition performance is better, but also suitable for data fusion.
Keywordscanonical correlation analysis feature fusion three-dimensional
Unable to display preview. Download preview PDF.
- 4.Pengfei, Y., Dan, X., Hao, Z.: Feature level fusion using palmprint and finger geometry based on Canonical Correlation Analysis. In: 3rd International Conference on Advanced Computer Theory and Engineering, pp. 260–264. ASME Press, Chengdu (2010)Google Scholar
- 5.Jing, X., Li, S., Lan, C., Zhang, D., Yang, J., Liu, Q.: Color image canonical correlation analysis for face feature extraction and recognition. Signal Processing (2011) (in press, corrected proof)Google Scholar