Direct Motion Models for Assessing Generated Videos
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915268141252608 |
|---|---|
| author | Allen, Kelsey Doersch, Carl Zhou, Guangyao Suhail, Mohammed Driess, Danny Rocco, Ignacio Rubanova, Yulia Kipf, Thomas Sajjadi, Mehdi S. M. Murphy, Kevin Carreira, Joao van Steenkiste, Sjoerd |
| author_facet | Allen, Kelsey Doersch, Carl Zhou, Guangyao Suhail, Mohammed Driess, Danny Rocco, Ignacio Rubanova, Yulia Kipf, Thomas Sajjadi, Mehdi S. M. Murphy, Kevin Carreira, Joao van Steenkiste, Sjoerd |
| contents | A current limitation of video generative video models is that they generate plausible looking frames, but poor motion -- an issue that is not well captured by FVD and other popular methods for evaluating generated videos. Here we go beyond FVD by developing a metric which better measures plausible object interactions and motion. Our novel approach is based on auto-encoding point tracks and yields motion features that can be used to not only compare distributions of videos (as few as one generated and one ground truth, or as many as two datasets), but also for evaluating motion of single videos. We show that using point tracks instead of pixel reconstruction or action recognition features results in a metric which is markedly more sensitive to temporal distortions in synthetic data, and can predict human evaluations of temporal consistency and realism in generated videos obtained from open-source models better than a wide range of alternatives. We also show that by using a point track representation, we can spatiotemporally localize generative video inconsistencies, providing extra interpretability of generated video errors relative to prior work. An overview of the results and link to the code can be found on the project page: http://trajan-paper.github.io. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_00209 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Direct Motion Models for Assessing Generated Videos Allen, Kelsey Doersch, Carl Zhou, Guangyao Suhail, Mohammed Driess, Danny Rocco, Ignacio Rubanova, Yulia Kipf, Thomas Sajjadi, Mehdi S. M. Murphy, Kevin Carreira, Joao van Steenkiste, Sjoerd Computer Vision and Pattern Recognition Machine Learning A current limitation of video generative video models is that they generate plausible looking frames, but poor motion -- an issue that is not well captured by FVD and other popular methods for evaluating generated videos. Here we go beyond FVD by developing a metric which better measures plausible object interactions and motion. Our novel approach is based on auto-encoding point tracks and yields motion features that can be used to not only compare distributions of videos (as few as one generated and one ground truth, or as many as two datasets), but also for evaluating motion of single videos. We show that using point tracks instead of pixel reconstruction or action recognition features results in a metric which is markedly more sensitive to temporal distortions in synthetic data, and can predict human evaluations of temporal consistency and realism in generated videos obtained from open-source models better than a wide range of alternatives. We also show that by using a point track representation, we can spatiotemporally localize generative video inconsistencies, providing extra interpretability of generated video errors relative to prior work. An overview of the results and link to the code can be found on the project page: http://trajan-paper.github.io. |
| title | Direct Motion Models for Assessing Generated Videos |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2505.00209 |