Content-based image segmentation

This paper explains how to extract feature points from an image and then segment the image based on these features. The feature extraction algorithms are robust enough since they are invariant to geometrical transformations including rotation, stretch and skew. The feature points are used to define Voronoi cells and image segments based on the Voronoi diagram. These segments can be used for a number of potential applications including image compression and image watermarking. This will enhance regional based algorithms.