MRF model with implicit shape prior for object shape segmentation
Sun Jian-zhon · Jisuanji gongcheng yu sheji · 2014
To deal with shape segmentation of the object in complicated scenes,a novel method based on MRF model with implicit shape prior is proposed.An implicit shape model with level set signed distance function is built to improve the accuracy of object segmentation,and a MRF energy function with the constraint of the existing shape prior is constructed.The optimal value of energy function is found by graph cut algorithm.The precise segmentation of object is obtained by applying the shape alignment and max-flow method to evolve the initial contour.Experimental results show that the proposed method can better cope with the clutter and noise,as well as partial occlusions and affine transformation of the shape,thus the robust stability of segmentation result is improved.