Can dense 2D supervision help learning 4D?
Introducing UFO-4D, feedforward dense 4D reconstruction from two unposed images. By leveraging 2D signals to bridge sparse 4D datasets, we enable high-fidelity 4D interpolation in one forward pass.
๐ ufo-4d.github.io
#ICLR2026
Joined May 2017
- Highly insightful talk from Zachary Teed, introducing his work on optical/scene flow, stereo, and the very recent work DROID-SLAM!
- Happy to be listed as one of the outstanding reviewers, three years in a row from 2018! I will keep trying to contribute better and better every years ๐Always fun to see one's name in list of helpful reviewers. Thanks @CVPR. Good to also see familiar twitter friends and colleagues @jan_gemert @abursuc @anfurnari @SamuelAlbanie @DocXavi, Tushar Nagarajan, Jean-Baptiste Alayrac. Sharing any list with AZ @Oxford_VGG is an honour


