Multimodal semantic segmentation of depth and RGB images combines the rich colour cues of conventional cameras with the geometric insight provided by depth sensors to partition a visual scene into ...