The Research On Medical Image Segmentation Algorithm Based On Level Set

Image segmentation is an important technology in the fields of image processing,computer vision,and medical image analysis.This technology is widely used in industry,agriculture,transportation,medicine and other fields.In the medical field,since medical diagnosis and treatment instruments are susceptible to interference from the external environment during the imaging process,it is difficult to obtain medical images with clear boundaries and uniform gray levels.Therefore,image segmentation is still very challenging for such problems.In recent years,image segmentation methods based on level sets have developed rapidly.It uses smooth closed contours as the final segmentation result,and can obtain accurate target boundaries,which can be adapted to imaging complex medical images.This thesis is based on the level set algorithm and the main work is as follows:1.A level set segmentation model based on guided filtering and fuzzy clustering.This method combines level set and guided filter fuzzy clustering algorithm to segment medical images.Because the guided filter uses the image itself and the spatial information of the guided image to change the pixel value,preserves some edge details of the image and can suppress a certain density of noise,the image can be approximated by pre-segmenting the image through the guided filter fuzzy clustering algorithm Obtain the segmentation target,thereby obtain the control parameters of the level set evolution,realize the automation of the initial contour of the level set,and avoid the influence of manual intervention.At the same time,due to the edge retention and noise suppression of the guided filter,the model can obtain more regional boundary information and has a certain robustness to noise.2.A level set segmentation model based on local area fitting information and guided filter fuzzy clustering.This method proposes a variable regional energy coefficient,which has a larger value near the smooth area and a smaller value at the target edge,so as to ensure that the level set evolution has a faster convergence speed in the smooth area and converges near the target edge.The speed is slower and it is not easy to lose the target edge.At the same time,combined with the guided filter fuzzy clustering algorithm,the influence of human factors is reduced,and the Laplacian of Gaussian energy item is introduced to enhance the edge information of the target.Compared with other models,the improved model is more robust to the initial contour and noise,and the segmentation accuracy is improved.3.A medical image segmentation algorithm platform based on level sets.In order to verify the feasibility and effectiveness of this algorithm and deepen the understanding of image segmentation algorithms,this thesis uses Matlab development tools to develop a level set-based medical image segmentation algorithm platform.The platform can provide users with simple interactive operations,demonstrate the curve iteration process when running the algorithm,and then compare it with different models,and use the similarity criterion to obtain quantitative data for algorithm evaluation.The experimental results show that the improved method proposed in this thesis can effectively segment and image complex medical images,and has a good segmentation effect on images with low noise.In chapter three,the model is robust to the initial contour and noise to a certain extent.In chapter four,the model is not only more robust to the initial contour and noise,but also can obtain more accurate target edges and achieve higher segmentation accuracy.