RGB-D Semantic Segmentation: A Review

Abstract

Semantic segmentation aims at giving a class label for each pixel on the image according to its semantic meaning. This problem is one of the most challenging tasks in computer vision, and has received a lot of attention from the computer vision community. In particular, with the popularity of depth cameras, many researchers have used rich structural information of depth data to assist semantic segmentation. In this paper, we have conducted a concise review of RGB-D semantic segmentation methods, including traditional method and deep learning method. In addition, discussions on RGB-D semantic segmentation are also provided.