Unsupervised segmentation of stereoscopic video objects: investigation of two depth-based approaches

Abstract

In this paper two efficient unsupervised video object segmentation approaches are proposed and thoroughly compared in terms of computational cost and quality of segmentation results. Both methods are based on the exploitation of depth information, estimated for stereoscopic pairs of frames. In particular, in both schemes an occlusion compensated disparity field is initially computed and a depth map is generated. Then a depth segments map is produced by incorporating a modified version of the multiresolution Recursive Shortest Spanning Tree segmentation algorithm (M-RSST). Next considering the first "Constrained Fusion

Modified Fuzzy C-Means as a Stereo Segmentation Method

“…The underlying idea of our algorithm is to consider the disparity space (e.g., in disparity maps) as a specific type of the data set, consisting of clusters representing the three dimensional objects of the scene. The fuzzy c-means algorithm has already been used to create the segmentations based on the depth information or disparity maps, e.g., (Ntalianis et al, 2002;Aik and Choon, 2011), and was also adapted to incorporated the spatial neighbourhood information, e.g., (Liew et al, 2000;Chuang et al, 2006;Meena and Raja, 2013), but in all these approaches, the algorithms were run on the input data already containing the depth information for each processed point. Our algorithm does not need the depth information in advance, since it calculates it itself by means of the stereo matching.…”