Topographic Independent Component Analysis Based on Fractal and Morphology Applied to Texture Segmentation

Abstract

The topographic independent component analysis (TICA) is a technique for texture segmentation in which the image base is obtained from the mixture matrix of the model through a bank of statistical filters. The use of energy as the topographic criterion in connection with the TICA filter bank has been explored in the literature, with good results. In such context, this paper proposes the use of energy plus a morphologic fractal descriptor as a new topographic criterion to be used in connection with the TICA filter bank. The new approach, called TICA fractal multi-scale (TICAFS) approach, results in a meaningful reduction of the segmentation error and/or in a meaningful reduction in the number of filters, when compared to the TICA energy (TICAE) approach.