Learning Common Representation from RGB and Depth Images

Giorgio Giannone, Boris Chidlovskii

We propose a new deep learning architecture for the tasks of semantic segmentation and depth prediction from RGB-D images. We revise the state of art based on the RGB and depth fea…