An Improved Image Segmentation Algorithm Based on Local Region Active Contour Model

Abstract

Taking into account the problem that the intensity in homogeneity effect of the image segmentation results, this paper proposes a new signal pressure function. First of all, Edge-based GAC (Geodesic Active Contours) model and region-based C-V (Chan-Vese) model are introduced, then SBGFRLS model (Selective Binary and Gaussian Filtering Regularized Level Set model)is being extracted, and a new signed pressure force (SPF) function is used, so that the method can deal with the image of intensity in homogeneity. The proposed method pays more attention to the details of the image, getting a more accurate segmentation results.