Image Segmentation Based on Integrated Gradient Vector Field

Abstract

GAC (geodesic active contour) model is the most basic and important image segmentation method based on PDE (partial differential equation), but the smoothness of images has a notable influence on the segmentation result of images, so how to choose a proper smoothing scale is of importance in researches. In this paper, we proposed an integrated gradient vector field instead of single-scaled gradient vector of images, to improve the performance of GAC model. The experimental results show that our method has a faster convergent speed and a better anti-noise ability than the traditional GAC model