- Submitted by
- Gengshan Yang
- Authors
- Gengshan Yang and Deva Ramanan
- Venue
- NeurIPS
- Abstract
- Gengshan Yang and Deva Ramanan. Volumetric Correspondence Networks for Optical Flow, NeurIPS 2019. Submission to RVC.
Images are downsampled to 0.6x resolution.
Multiple submissions of the same result due to delayed server response.
- Runtime
- 700 ms
- External data used
- FlyingChairs, FlyingThings3D, Sintel, KITTI, HD1k, Viper, generated GTAV data
- Links
-
- Number of submissions
- 3
- Detailed results
-
| WAUC (day) |
WAUC (sunset) |
WAUC (rain) |
WAUC (snow) |
WAUC (night) |
Mean WAUC over conditions |
| 71.0 |
62.2 |
53.9 |
61.4 |
56.0 |
60.9 |