Datasets that feed robot vision and SLAM.
6,000 real scanned objects with meshes, videos and point clouds.
