Multifractal analysis of gray-scale images

The basis of fractal (and multifractal) theory and its application to image analysis is presented. In the multifractal approach the image is described by a measure /spl mu/, or by capacities. Through the multifractal spectrum f(/spl alpha/) both local and global information of the image regularity can be derived. By using a different fractal dimension different parts of analyzed images can be detected, without loosing or introducing any information. This feature is very useful for a global image analysis, edge detection, segmentation, texture classification, etc. The applicability of the multifractal approach to image analysis is verified through different medical images.