An adaptable active contour model for medical image segmentation based on region and edge information

Abstract

Due to the complexity of the internal structure of human body and the physiological movement of the illuminated tissue, the digital medical image exists low contrast, high noise intensity, complex internal structure and edge blur phenomenon, which will limit the segmentation accuracy of traditional active contour models. To solve this problem, this paper proposes a novel active contour model based on the combination of regional information and the edge information of the image. The new approach has four key characteristics. First the local information fitting of the image is incorporated into the pressure force function (SPF) of the SBGFRLS model, which improves the ability of dealing with medical images with low contrast and complex structure. Second, the adaptive balance of local information and global information is realized by adding a novel weighting function, which accelerates the evolution speed and enhances the adaptability of the model; Third, in the numerical implementation process of the proposed model, the divergence operator is replaced by the Gaussian filter, in this way, the level set function is smoothed and the computation is simplified. Last, a penalty term of symbolic function is introduced to reduce the computational complexity of the level set function due to re-initialization and regularization process. In order to verify the effectiveness of the model, we use different kinds of medical images for simulation experiments. Experimental results show that compared with the traditional active contour models, the proposed method can achieve an satisfactory both in segmentation speed and accuracy.