Text-to-Video Retrieval on MSR-VTT (Official full-size test)
48.8R@1CAMoE
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CAMoE2021.09 | 48.8 | 75.6 | — | — | — | — | |
| CoCaZero-shot=true, subset=true2022.05 | 30 | 52.4 | 61.6 | — | — | — | |
| CLIPZero-shot=true, subset=true2022.05 | 23.3 | 44.2 | 53.6 | — | — | — | |
| CLIPZero-shot=true, subset=false2022.05 | 21.4 | 41.1 | 50.4 | — | — | — | |
| AMMTrained On=S-MiT2021.05 | 17.3 | 39.8 | 51.8 | — | 28.4 | — | |
| Dual EncodingFeatures=Transformer features included2020.09 | 13.3 | 34 | 45.7 | 13 | 23.8 | — | |
| AMMTrained On=MSR-VTT2021.05 | 11.8 | 33.9 | 48.2 | — | 23.2 | — | |
| Dual EncodingSource=Proposed, Visual Feature=ResNeXt-ResNet2020.09 | 11.6 | 30.3 | 41.3 | 17 | 21.2 | 211.7 | |
| Dual EncodingFeatures=Standard 2D CNN2020.09 | 11.6 | 30.3 | 41.3 | 17 | 21.2 | — | |
| W2VV++Source=Re-trained, Visual Feature=ResNeXt-ResNet2020.09 | 11.1 | 29.6 | 40.5 | 18 | 20.6 | 191.4 | |
| HGRSource=Re-trained, Visual Feature=ResNeXt-ResNet2020.09 | 11.1 | 30.5 | 42.1 | 16 | 20.8 | 204.4 | |
| AMMTrained On=Vatex2021.05 | 10.3 | 28.7 | 39.3 | — | 19.8 | — | |
| CESource=Original Paper2020.09 | 10 | 29 | 41.2 | 16 | — | 191.9 | |
| TCESource=Re-trained, Visual Feature=ResNeXt-ResNet2020.09 | 9.3 | 27.3 | 38.6 | 19 | 18.7 | 177.3 | |
| HGRSource=Original Paper2020.09 | 9.2 | 26.2 | 36.5 | 24 | — | 172.4 | |
| VSE++Source=Re-trained, Visual Feature=ResNeXt-ResNet2020.09 | 8.7 | 24.3 | 34.1 | 28 | 16.9 | 167.9 | |
| Dual EncodingFeatures=Transformer features only2020.09 | 8.5 | 25.1 | 36.1 | 22 | 17.4 | — | |
| CESource=Re-trained, Visual Feature=ResNeXt-ResNet2020.09 | 7.9 | 23.6 | 34.6 | 23 | 16.5 | 155.1 | |
| TCESource=Original Paper2020.09 | 7.7 | 22.5 | 32.1 | 30 | — | — | |
| AMMTrained On=ActivityNet2021.05 | 7.5 | 22.1 | 31.2 | — | 15.4 | — | |
| Mithun et al.Source=Original Paper2020.09 | 7 | 20.9 | 29.7 | 38 | — | 144.6 | |
| MEESource=Re-trained, Visual Feature=ResNeXt-ResNet2020.09 | 6.8 | 20.7 | 31.1 | 28 | 14.7 | 148 | |
| Francis et al.Source=Original Paper2020.09 | 6.5 | 19.3 | 28 | 42 | — | — | |
| W2VVSource=Re-trained, Visual Feature=ResNeXt-ResNet2020.09 | 1.1 | 4.7 | 8.1 | 236 | 3.7 | 117.9 |