The Chan-Vese Algorithm

TL;DR

The paper discusses the Chan-Vese algorithm, a flexible active contour method based on the Mumford-Shah functional, for image segmentation, especially in medical imaging, with implementation details in MATLAB.

Contribution

It presents the Chan-Vese model for image segmentation, explaining its formulation, advantages over classical methods, and providing MATLAB implementation guidance.

Findings

Effective segmentation of complex images including medical images.

Flexible model capable of handling various image types.

Implementation in MATLAB facilitates practical application.

Abstract

Segmentation is the process of partitioning a digital image into multiple segments (sets of pixels). Such common segmentation tasks including segmenting written text or segmenting tumors from healthy brain tissue in an MRI image, etc. Chan-Vese model for active contours is a powerful and flexible method which is able to segment many types of images, including some that would be quite difficult to segment in means of "classical" segmentation - i.e., using thresholding or gradient based methods. This model is based on the Mumford-Shah functional for segmentation, and is used widely in the medical imaging field, especially for the segmentation of the brain, heart and trachea. The model is based on an energy minimization problem, which can be reformulated in the level set formulation, leading to an easier way to solve the problem. In this project, the model will be presented (there is an extension to color (vector-valued) images, but it will not be considered here), and Matlab code that implements it will be introduced.