Next Previous Contents

1. Overview

The Segmentation menu is currently composed of two items. As their names indicate, the first one, 1D signals WSA based segmentation deals with 1D signals, as the second one, Image Multifractal Segmentation allows to process images. However, the main difference between these two sub-menus lies in the methods they use. The first one is model-based, and relies on a generalization of IFS-s called Weakly Self Affine functions or SGIFS (note that you may synthesize directly Weakly Self Affine functions in the Synthesis menu, under Functions/Stochastic/SGIFS). In fact this is really a modeling method (as such, it should appear more appropriately in a Modeling menu, which does not exist yet), used here for segmentation purposes. The second approach is fully non parametric, and analyses the image through various features of its multifractal spectrum. Although it is perfectly possible to use WSA functions to model and segment images as well as to use multifractal tools for the segmentation of 1D signals, these facilities are not implemented in the current version of Fraclab.


Next Previous Contents