Color and Grey Level Image Indexing Based on Tridimensional Representation of Force Histogram
Abstract
A new method for image indexing based on force histogram notion is proposed in this paper. This kind of histogram has low time computing and allows to keep fundamental geometric transformations (scale, translation and rotation). We propose to extend this notion to both grey level and color objects by adding a third dimension to take into account photometric variations. Experimental results on three different databases of images show the promising aspect of our approach.