Abstract: This paper aims at high-accuracy 3D object detection in autonomous driving scenario. We propose Multi-View 3D networks (MV3D), a sensory-fusion framework that takes both LIDAR point cloud ...
Abstract: Automated 3D reconstruction of faces from images is challenging if the image material is difficult in terms of pose, lighting, occlusions and facial expressions, and if the initial 2D ...