Depth Image Generation from monocular RGB image

Abstract

In recent years, Visual SLAM using only monocular RGB camera has been researched. In this paper, we propose a method to generate depth images from a single RGB image using GAN, which is one of deep learning. As a result of the verification experiments, it is suggested that the depth prediction can be done, although it is large for both indoor and outdoor images. Future tasks include verification experiments using large-scale datasets.