Image segmentation with global active contours
A. Volkan Atli, İlker Bayram · 2012
In this paper, a new segmentation method which is derived from distance regularized level set evolution (drlse) equations is proposed. The proposed method merges drlse equations and a signed pressure force function. Thus, a new global active contour model which is gradient and also region-based is obtained from the only gradient based drlse method. In addition, the method is also independent from the location of initial contour and it can extract regions of interest inside or outside the contour. Finally, the proposed method with various active contour methods, tested on real and synthetic images, the results were compared.