Image Segmentation Algorithm Based On Level Set And Its Application To Medical Images

Image segmentation is an important and basic research content in image processing,and is widely used in many fields.In the medical field,image segmentation often plays a key role in the medical research problems and clinical application.The difficulty of image segmentation is that there is no universally applicable method.In different applications,because of different targets of interest and different types of images,different segmentation methods are often proposed.In order to obtain satisfactory segmentation results,many excellent segmentation algorithms have been proposed.Among them,the level set algorithm has shown significant advantages in image segmentation problems.Under this background,this dissertation introduces the level set algorithm and proposes two level set models for medical images.The main contents of our studies are summarized as follows:1)Study the image segmentation methods in recent years,conduct classification analysis and understanding,then we focus on the basic principles of level set segmentation algorithm,introduce the classic level set model and analyze the deficiencies;2)Aiming at the difficulty of segmentation of medical images,based on the classic level set model,a level set segmentation model combining global and local information is proposed,named HLSGL model,and a new iteration stop condition is designed.The proposed model can be used in different types of medical images to obtain good segmentation results,and compared with other level set models to verify the superiority of the proposed model;3)For the cyst images in medical imaging,a local fitting C-V level set model is proposed,named LFCV model.By using the local fitting image related to the Gaussian kernel and introducing local area information,the model has certain robustness on the intensity inhomogeneity.Experiments verified that the model has good performance in different types of cyst image segmentation.