MiDaS
Monocular Depth
- Submitted by
- Stephan Richter
- Authors
- René Ranftl, Katrin Lasinger, David Hafner, Konrad Schindler, Vladlen Koltun
- Venue
- arXiv:1907.01341
- Abstract
- Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer. Standard pretrained model from Dec 2019 (see code link).
- Runtime
- 100 ms
- External data used
- Default MiDaS model, not trained on VIPER
- Links
- Number of submissions
- 1
- Detailed results
-
Metric All day sunset snow rain night Scale invariant logarithmic error 0.89 0.73 0.84 0.88 1.08 0.93 Absolute squared error [%] 21.89 11.93 12.49 29.75 27.89 27.40 Root mean squared error of the inverse depth 45.18 20.61 26.36 50.61 74.43 53.90