The authors built a unified model with two operating modes.
Last-frame layouts guide video generation through large camera moves
LIFT combines a camera trajectory with a target layout for the final frame, allowing users to specify content in regions initially outside the camera’s view.
Big Tech
Shengxiang Ji · Boyang Wang · Haiyang Xu · Bingnan Li · Yucheng Mao · Zeyuan Chen · +6 more
UC San Diego · University of Virginia · Meta · Amazon · Lambda
Research Digest··3 min read
Ji and colleagues introduce LIFT, an image-to-video system designed for camera movements that reveal substantial unseen areas.
Why this paper
From Amazon and 4 others
In one line
LIFT enables specifying layout of future views in video generation via on-policy self-distillation from a dense-layout teacher, improving controllability under large viewpoint changes.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§