Point Tracking on TAP-Vid DAVIS (Strided)
53.1Avg Delta ErrorTAP-Net
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| TAP-NetTraining Data=Kubric2023.07 | 53.1 | 38.4 | 82.3 | |
| TAP-NetEvaluation Mode=Strided, Input Resolution=256x2562024.07 | 53.1 | 38.4 | 82.3 | |
| TAPNetTraining Dataset=Kub, Evaluation Resolution=256 × 2562025.07 | 53.1 | 38.4 | 82.3 | |
| PIPsEvaluation Mode=Strided, Input Resolution=256x2562024.07 | 59.4 | 42 | 82.1 | |
| Context-PIPsEvaluation Mode=Strided, Input Resolution=256x2562024.07 | 64 | 48.9 | — | |
| OmniMotionTraining Data=Not Specified2023.07 | 67.5 | 51.7 | 85.3 | |
| OmniMotionEvaluation Resolution=256 × 2562025.07 | 67.5 | 51.7 | 85.3 | |
| PIPsTraining Data=FlyingThings++2023.07 | 70 | 52.4 | 83.6 | |
| PIPsTraining Dataset=FT [42], Evaluation Resolution=384 × 5122025.07 | 70 | 52.4 | 83.6 | |
| MFTTraining Data=Kubric+more2023.07 | 70.8 | 56.1 | 86.9 | |
| MFTEvaluation Mode=Strided, Input Resolution=256x2562024.07 | 70.8 | 56.1 | 86.9 | |
| TAPIRTraining Data=Kubric2023.07 | 73.6 | 61.3 | 88.8 | |
| TAPIREvaluation Mode=Strided, Input Resolution=256x2562024.07 | 73.6 | 61.3 | 88.8 | |
| TAPIRTraining Dataset=Kub, Evaluation Resolution=256 × 2562025.07 | 73.6 | 61.3 | 88.8 | |
| PIPs++Training Data=Point Odyssey2023.07 | 73.7 | — | — | |
| Dino-TrackerEvaluation Resolution=256 × 2562025.07 | 78.2 | 62.3 | 87.5 | |
| BootsTAP+Evaluation Mode=Strided, Input Resolution=256x256, Extra Training Data=15M video clips2024.07 | 78.5 | 66.4 | 90.7 | |
| BootsTAPIRTraining Dataset=Kub+15M, Evaluation Resolution=256 × 2562025.07 | 78.5 | 66.2 | 90.7 | |
| TAPTRv2Evaluation Mode=Strided, Input Resolution=256x256, Backbone=ResNet-502024.07 | 78.8 | 66.4 | 91.3 | |
| TAPTRv2Training Dataset=Kub, Evaluation Resolution=256 × 2562025.07 | 78.8 | 66.4 | 91.3 | |
| CoTracker-SingleEvaluation Mode=Strided, Input Resolution=256x2562024.07 | 79.1 | 64.8 | 88.7 | |
| TAPTREvaluation Mode=Strided, Input Resolution=256x256, Backbone=ResNet-502024.07 | 79.2 | 66.3 | 91 | |
| TAPTRTraining Dataset=Kub, Evaluation Resolution=256 × 2562025.07 | 79.2 | 66.3 | 91 | |
| CoTrackerTraining Data=Kubric2023.07 | 79.4 | 65.9 | 89.9 | |
| CoTracker2-SingleEvaluation Mode=Strided, Input Resolution=256x2562024.07 | 79.4 | 65.9 | 89.9 | |
| CoTracker2Training Dataset=Kub, Evaluation Resolution=384 × 5122025.07 | 79.4 | 65.9 | 89.9 | |
| LocoTrackTraining Dataset=Kub, Evaluation Resolution=256 × 2562025.07 | 79.6 | 67.8 | 89.9 | |
| TAPNextTraining Dataset=Kub, Evaluation Resolution=256 × 2562025.07 | 79.7 | 65.4 | 88.9 | |
| Anthro-LocoTrackTraining Dataset=Kub+1.4K, Evaluation Resolution=256 × 2562025.07 | 81 | 69 | 90.8 | |
| LocoTrackTraining Dataset=Kub, Evaluation Resolution=384 × 5122025.07 | 81.3 | 69.4 | 88.6 | |
| BootsTAPNextTraining Dataset=Kub+15M, Evaluation Resolution=256 × 2562025.07 | 82.4 | 68.9 | 91.6 | |
| Anthro-LocoTrackTraining Dataset=Kub+1.4K, Evaluation Resolution=384 × 5122025.07 | 82.9 | 71.1 | 90.3 | |
| Anthro-TAPNextTraining Dataset=Kub+1.4K, Evaluation Resolution=256 × 2562025.07 | 83.5 | 71.4 | 92.4 |