next up previous contents
Next: From restoration to segmentation Up: Image processing Previous: Extension to unsupervised classification   Contents

Segmentation

This section is concerned with image segmentation, which aim is to find a partition of an image into its constituent parts. The idea is still to apply our topological gradient based algorithm for the detection of edges in the image.

Several approaches have been studied in the literature. One can cite variational methods, for example based on the minimization of the Mumford-Shah functional [86], the active contours and snake methods [37,99], stochastic approaches [36,42], wavelets, ...[13,14,12,88,94,95,115].

This section summarizes the main results presented in [18,27]. Several numerical experiments are also detailed in these references and show the efficiency of our approach.



Subsections

Back to home page