Image Segmentation Based On Level Set

As an important part of image processing,image segmentation can be applied in many fields such as medical image analysis and product monitoring.There are many methods for image segmentation.Image segmentation model based on level set method is a subject that scholars have studied more in recent years.Because it has the advantages of flexible processing of image topology changes and simple numerical calculation,it is more and more widely used in the field of image segmentation.However,there are still many problems in the image segmentation model based on the level set method,such as low accuracy and precision of segmentation,sensitivity to initial contour,etc.As a result,we can not get the desired segmentation results,so the theory and model still need to be improved accordingly.In view of the advantages and disadvantages of the existing models,this paper constructs new segmentation models,which mainly includes the following improvements:(1)Based on the edgeless active contour(CV)model,the piecewise constant fast segmentation model and local information-based(LCV)model,a fast segmentation model,LKCV model,is established to fuse local information of image.Compared with the previous model,LKCV model has the advantages of better segmentation of intensity inhomogeneous images and fewer iterations.(2)In order to improve the weakness of the local binary fitting(LBF)model,which has insufficient evolutionary power and is sensitive to the initial contour,a regularized area term is added to the LBF model,and an area term coefficient which can change automatically according to the gray level information of the image is defined,a local binary fitting model with a variable coefficient regularized area term,ALBF model,is obtained.(3)In order to alleviate the shortcomings of the slow evolution of the proposed ALBF model,the global term of CV model is added to ALBF model in this paper,and the edge-preserving fast segmentation model based on global and local information,GLBF model,is obtained.This model can not only segment the image with intensity inhomogeneity,but also process the image with weak boundary and deep depression better.At the same time,the evolution time of GLBF model is less than that of ALBF model.