notesum.ai
Published at December 6Mind the Time: Temporally-Controlled Multi-Event Video Generation
cs.CV
Released Date: December 6, 2024
Authors: Ziyi Wu1, Aliaksandr Siarohin, Willi Menapace, Ivan Skorokhodov, Yuwei Fang, Varnith Chordia, Igor Gilitschenski, Sergey Tulyakov
Aff.: 1Snap Research, University of Toronto, Vector Institute
![[Uncaptioned image]](https://arxiv.org/html/2412.05263v1/x1.png)
| Method | Subject | Background | Aesthetic | Imaging | Motion | Dynamic |
|---|---|---|---|---|---|---|
| Consist. | Consist. | Quality | Quality | Smooth | Degree | |
| Short | 0.857 | 0.939 | 0.498 | 0.583 | 0.995 | 0.481 |
| Global | 0.890 | 0.950 | 0.541 | 0.613 | 0.995 | 0.517 |
| Ours | 0.900 | 0.950 | 0.544 | 0.609 | 0.988 | 0.711 |