Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-scale Convolutional Architecture
In this paper we address three different computer vision tasks using a single basic architecture: depth prediction, surface normal estimation, and semantic labeling. We use a mu...