A Level Set Algorithm Based on Probabilistic Statistics for MR Image Segmentation

MR image segmentation is of great importance in medical image application. MR images have the characteristics of intensity inhomogeneities, strong background interference and blurred target area. These characteristics will greatly increase the difficulty of segmentation and affect image segmentation results. To obtain the satisfied performance of MR image segmentation, a level set algorithm based on probabilistic statistics for MR image segmentation is proposed. Because of the intensity inhomogeneity of the image, a bias field is used to describe the image in the proposed model. But the addition of a bias field will increase the amount of computation. Therefore, combining with the probabilistic statistical theory, the energy function is defined by the pixel distribution probability to improve operational efficiency. In addition, a new rule item is added to enhance the edge information of the image to highlight the edge segmentation curve. Experimental results show that the proposed model behaves well in segmenting MR images.