Segmentation of Structures in Medical Images: Review and a New Computational Framework

This paper aims to make a brief survey on the current segmentation algorithms used for medical images. Instead of describing the detailed procedures, we will summarize their main ideas, enumerate their merits and drawbacks and discuss their possible applications. The algorithms are classified into three main categories: threshold-based models, pattern recognition models and deformable models. Due to the appreciable advantages of deformable models and their promising applications, we mainly focus on the discussion on deformable models. In the second part of this paper, we describe a new computational framework developed for medical image processing.