I'm curious if stereo vision and eyepoint movement has a benefit for humans in distinguishing shape vs texture, and perhaps that could be a promising approach for ML, but are there any databases for that? There are car focused ones like KITTI, but any more generalist ones?