Optical Flow Estimation On Kitti 2015 Train
Métriques
EPE
F1-all
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | EPE | F1-all |
|---|---|---|
| volumetric-correspondence-networks-for | 8.36 | 25.1 |
| craft-cross-attentional-flow-transformer-for | 4.88 | 17.5 |
| raft-recurrent-all-pairs-field-transforms-for | 5.04 | 17.4 |
| pwc-net-cnns-for-optical-flow-using-pyramid | 10.35 | 33.7 |
| fastflownet-a-lightweight-network-for-fast | 12.24 | 33.1 |
| learning-to-estimate-hidden-motions-with | 4.69 | 17.1 |
| flownet-20-evolution-of-optical-flow | 10.08 | 30.0 |
| flowformer-a-transformer-architecture-for | 4.09 | 14.7 |
| hierarchical-discrete-distribution | 13.17 | 24.0 |
| rethinking-raft-for-efficient-optical-flow | 4.83 | 16.45 |
| recurrent-partial-kernel-network-for | 3.79 | 13.0 |
| rapidflow-recurrent-adaptable-pyramids-with | 5.87 | 17.7 |
| deep-equilibrium-optical-flow-estimation | 3.76 | 13.0 |
| global-matching-with-overlapping-attention | 4.24 | 15.4 |
| maskflownet-asymmetric-feature-matching-with | - | 23.1 |
| learning-optical-flow-from-a-few-matches | 6.80 | 19.3 |
| separable-flow-learning-motion-cost-volumes | 4.60 | 15.9 |
| dpflow-adaptive-optical-flow-estimation-with-1 | 3.37 | 11.1 |