Image Segmentation Based On Active Contour Model And Level Set Method

Image segmentation technology is a key technology in image processing, and it gives a direct influence on image analysis and understanding. Being capable of detecting the snatchy boundaries by a closed curve, active contour models (ACM) have received a wide attention. Level set method successfully deals with the topological change of active contour in an elegant way by evolving a level set function defined in a higher dimensional space instead of the curve represented by parametric equation. ACM and level set methods have been an important topic in image segmentation field.In order to provide an open platform for medical image segmentation algorithms, one active contour model and two level set methods are analyzed and compared. With some additional image common operations, the open algorithm platform is built, and integrates these three methods encoded in dynamic link library. The author firstly implements and compares GVF-Snake with the typical active contour model, i.e., snake; secondly, level set without reinitialization (LSWR) and level set without edges (LSWE) are implemented and analyzed; finally, presents the open algorihtm platform developed by using Microsoft MFC. The platform possesses the basic image operations, could handle medical images encoded by DICOM format and support some popular image format, e.g., BMP and JPEG. Meanwhile, it provides an open interface to integrate the aforementioned three algorithms, by which the programs are shared on the binary code level.