Research On Image Segmentation Method Based On Parameter Active Contour Model

Image segmentation is a basic problem in the field of image processing,also a key step in image processing and analysis.On the one hand it is the basis of object expression,directly affect the characteristics of the measuring target;on the other hand,image segmentation and target separation based on segmentation,feature extraction and parameter measurement transfer the original image into more abstract,more simple form of expression,which makes the image more easy to understand and analyze.Many kinds of image segmentation methods have been proposed,but because of the diversity and complexity of the target image,image segmentation is still an important and challenging research topic.In all image segmentation methods,the active contour model is a kind of method which is of great concern.It has rigorous mathematical theory foundation,and makes full use of the idea of dynamic evolution of contour curve,which is difficult to solve the problem of much excellent image segmentation,and fully reflects the superiority of the method.Based on the research of active contour model,found the true boundary segmentation image approximation of the initial contour must be the target of interest,otherwise it may be not perfect segmentation results,and because the non-convex active contour model,the cavity boundary which is difficult to approximate the target of interest.This paper studies an improved parametric active contour model,to improve the defects of the above problems,the model of the transformation of external energy functional to achieve,the external force is the external force field and vector field convolution integration force of the original method,finally get a global force.In the global force and external force,the elastic deformation of the template itself,and the local image feature matching to achieve harmony,is a function of energy minimization,to complete the image segmentation.Through the simulation experiment,and compared with other similar methods,this method can accurately segment the image target,and has a certain degree of noise immunity.