Two-dimensional CLAFIC methods for image recognition

Abstract

In this paper, we propose two variations of vector based class-featuring information compression (CLAFIC) methods which can be applied directly to the gray level digital image data. In these methods, gray level digital image matrix data is processed without any explicit transformation into the vector form. Therefore, we called them as two-dimensional CLAFIC methods. Evaluation of correlation and covariance matrices from the matrix forms of the image data speeds up the training and test phases of image recognition applications. Experimental results on the AR and the ORL face databases demonstrate that the proposed two-dimensional CLAFIC methods are more efficient than the conventional CLAFIC and some other methods given in the paper.