“…Most of the approaches have used Principle Component Analysis (PCA) for building efficient representations and for subsequent recognition. The approach has led to a variety of successful applications, e.g., human face recognition [22,2], visual inspection [25], visual positioning and tracking of robot manipulators [15], illumination planning [14], mobile robot localization [11], and background modeling [17]. However, the standard way to perform recognition, based on projections, is prone to errors in the case of non-Gaussian noise, e.g., occlusions, varying illumination conditions, and cluttered background in the input images.…”