Improved Compound Vector Field Based Active Contours Model

Abstract

The Active Contours Model (ACM) is an active researching area in medical image segmentation. In traditional ACM model, boundary of region of interest (ROI) can be obtained by deforming the spline curve. But the segmentation relies on the initial location of the curve which is apt to be converged to the local gradient maximum region. Moreover, the model cannot segment the concave region accurately. In this paper, an improved ACM model algorithm for image segmentation based on the compound vector field is proposed. The image is processed by the generalized fuzzy theory to get a better edge map. The improved ACM model achieves a better effect on image segmentation by replacing the traditional GVF (gradient vector flow) with the compound vector field. The segmentation experiments show that the algorithm has brilliant capacity of not only capturing the image feature in a wider region but also dealing with the concave regions. Index Terms—ACM model, generalized fuzzy, compound vec- tor field.