A Review of Medical Image Segmentation: Methods and Available Software

Automatic medical image segmentation is an unsolved problem that has captured the attention of many researchers. The purpose of this survey is to identify a representative set of methods that have been used for automatic medical image segmentation over the past 35 years and to provide an opportunity to view the transitions that have occurred as this research area has developed. To facilitate this, the existing research is divided into three generations, each generation adding an additional level of algorithmic complexity. These generations indicate progress towards accurate, fully-automatic, medical image segmentation and their identification provides a framework for classifying the wide variety of methods that have been devised. The first generation is composed of the simplest forms of image analysis such as the application of intensity thresholds and region growing. The second generation is characterized by the application of uncertainty models and optimization methods, and the third generation incorporates knowledge into the segmentation process. The progress toward accurate, fully-automatic segmentation is discussed and sources of segmentation software from industry and academia are identified, along with databases for segmentation validation.