Depth-Wise Segmentation of 3D Images Using Entropy

Recent advances in 3D modeling and depth estimation of objects have created many opportunities for multimedia computing. Using depth information of a scene enables us to propose a brand new segmentation method called DepthWise segmentation. Unlike the conventional image segmentation problems which deal with surface-wise decomposition, the depth-wise segmentation is a problem of slicing an image containing 3D objects in a depth-wise sequence. The proposed method uses entropy of a depth image to characterize the edges of objects in a scene. Later, obtained edges are used to find Line-Segments. By linking the line-segments based on their object and layer numbers, ObjectsLayers are achieved. To test the proposed segmentation algorithm, we use syntactic images of some 3D scenes and their depth maps. The experiment results show that our method gives good performance.