Teaching 2D cameras to see 3D

Facebook AI researchers Georgia Gkioxari, Shubham Tulsiani, and David Novotny delivered a fascinating paper at the International Conference on Computer Vision (ICCV) in Seoul, titled Pushing state-of-the-art in 3D content understanding. The premise of the paper is that to interpret the world around us, AI systems must understand visual scenes in three dimensions. To accomplish that formable task, the researchers developed a novel technique, called VoteNet, to perform object detection for circumstances when 3D input from LIDAR or other sensors is available. While most traditional systems for this task depend on 2D image signals, theirs is based purely on 3D