A Level Set Segmentation Algorithm Based on Saliency Guidance and Convexity Preservation
Combined with the anatomical features of the optic nerve disc in color fundus images,the curvature change of the zero-level set curve is introduced to optimize the active contour model,and a horizontal set algorithm for fully automatic segmentation of the optic nerve disc is proposed.The algorithm uses the significance algorithm to quickly obtain the initial contour,and the initial contour is very close to the optic disc contour,which greatly improves the segmentation efficiency.In addition,the convex preservation mechanism can make the curve not be affected by the capillaries around the optic disc during the evolution process to maintain the convexity of the segmentation curve.Finally,the objective function is introduced in the iteration to evaluate the convergence of the algorithm.Experiments show that the proposed CRLSF algorithm is superior to the existing algo-rithm.
optic disclevel set methodconvexityimage segmentation
NETL
NSTL
万方数据
