Scale invariant texture classification with mathematical morphology

One of the most interesting fields of research in Digital Image Processing is the segmentation of an image into different objects. The detection of regions with different textures is one of the techniques used to achieve this objective. These techniques differ among each other in the texture parameters used and the way to obtain them. In this work the characteristics associated to each region will be obtained from the application of the Size Criteria of Successive Openings. This technique of the Mathematical Morphology analyses how the image interacts with different structuring elements. Taking parameters from the images, we obtain size distribution functions associated to each region. We demonstrate that we can obtain characteristics invariant under changes of scale using the statistical moments of the associated normalised density functions. These characteristics can be used as patterns for a further classification.