Segmentation Based On Regularized Level-set Method

Image segmentation is the key step for image analysis technique. Its main aim is to separate the whole image into several sub-regions which share one or more similar statistical properties. Image segmentation always restricts the development of high level image processing. In recent years the image segmentation methods based on level-set method have received extensive attention theoretically and technologically by the image processing experts, because it has very rigid mathematical base and efficient numerical scheme, in addition it is very promising to overcome some shortcomings of traditional image segmentation methods such as break points on final contour etc.In this paper, we first gave a comprehensive summary of present image segmentation methods. Then we surveyed the development of level-set method and introduced several famous level-set methods. After that we introduced the principle and solving programs of the level set algorithm in detail, and highlights the application of distance regularization level-set in image segmentation. Distance regularization level-set can avoid re-initialization.To overcome the defects of distance regularization level-set method working in segmentation of liver CT sequence, we proposed a distance regularization level-set segmentation method based on the prior information of liver. We first obtained early liver segmentation results with tow-step region growing. Then the early liver segmentation results were used as priori information to constraint the level set evolution, which could overcome the weak edges out of bounds efficiently. Between the upper and lower slices in a liver CT sequence, we used distance change to obtain the seed points for region growing automatically, which made the segmentation running automatically and avoided the error accumulation. Our algorithm is stable, and has higher accuracy.