Texture-driven parametric snakes for semi-automatic image segmentation

Anaïs Badoual, Michael Unser, Adrien Depeursinge

Abstract: Highlights•We present a texture-driven parametric snake for semi-automatic segmentation of a single structure in an image.•The framework allows for easy and friendly user interaction and can be trained on-the-fly from pixel collections provided by the user.•We propose a new energy functional that combines intensity and texture information.•We use Fisher’s linear discriminant analysis to balance the two types of image information.•We demonstrate the efficiency of the snake through an extensive validation on synthetic as well as on real data.