Range image segmentation and classiication using cooperative relaxational algorithm between H-K curvatures

평균 곡률과 가우시안 곡률의 상호 셥동 이완 알고리즘을 이용한 거리 영상의 분할과 분류

The range image is divided into surface regions which are homogeneous in their intrinsic properties. In this paper, we use cooperative relaxational algorithm between curvatures to escape local minima and choose optimal possibility to reserve edge. Cooperative relaxational algorithm between curvatures is relaxation process in which weights of center pixel;s and neighbor pixel's possiblility are determined adaptively by using deviation of curvatures. Experimental resutls show that the proposed method segments and classifies the range images more accurately compared to the other relational algorithms.