An Optimized Iteration Algorithm based on C-V Model and Graph Cuts

C-V model can self-adapt to the changes of curve topology but requires more iterations and needs more computing time. Graph cuts algorithm is good at getting the global optimum in a short time but not suitable for concave object extraction. To overcome the flaws of these two algorithms, an optimized iteration algorithm has been proposed. First the initial contour is deformed with an improved C-V model, which was without re-initialization during iterative process and the iteration stop condition is set by calculating changing area within the contour. Then the active contour is input to graph cuts algorithm. Dilates the contour into its neighborhood and formed an inner and an outer boundary seperatively, changes these two boundaries as source and sink, and obtains the final contour by graph cuts. Experiments show that this optimized algorithm reduces the iteration time, and has better effect and higher efficiency for image segmentation.