Two applications of RGB-D descriptors in computer vision

Abstract

In this paper an evaluation of RGB-D descriptors in the context of Object Recognition and Object Tracking is presented. Spinimages, CSHOT and ECV context descriptors were used for detecting objects in point clouds. Empirical evaluation over a dataset with ground truth shows that shape is the most important cue for RGB-D descriptors. However, texture helps discrimination when objects are large or have little structure.