Multi-Object Tracking on TAO (val)
57.5AssocAGTR
Evaluation Results
| Method | Links | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| GTRFPS=11.22022.03 | 57.5 | — | — | — | — | — | — | — | — | — | — | — | 22.5 | 45.8 | 36.8 | — | — | — | — | — | |
| GTRDetections=QDTrack2022.03 | 55.6 | — | — | — | — | — | — | — | — | — | — | — | 20.4 | 40.7 | 30.1 | — | — | — | — | — | |
| QDTrackFPS=5.42022.03 | 53.5 | — | — | — | — | — | — | — | — | — | — | — | 16.1 | 35.8 | 24.3 | — | — | — | — | — | |
| AEDDetector=Co-DETR2024.09 | 52.4 | 55.3 | 71.8 | 41.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HAT-MASA-G-DINODetection=Grounding DINO, Zero-shot transfer=true2025.03 | 46.7 | 47.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HAT-MASA-DeticDetection=Detic, Zero-shot transfer=true2025.03 | 46.4 | 47.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GLEE-ProPublication=CVPR20242024.09 | 46.2 | 47.2 | 66.2 | 29.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HAT-MASA-SAM-BDetection=SAM-B, Zero-shot transfer=true2025.03 | 45.6 | 46.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MASA-G-DINODetection=Grounding DINO, Zero-shot transfer=true2025.03 | 45 | 46.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MASA-DeticDetection=Detic, Zero-shot transfer=true2025.03 | 44.5 | 46.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HAT-MASA-R50Backbone=ResNet-50, Zero-shot transfer=true2025.03 | 44.4 | 46.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MASA-SAM-BDetection=SAM-B, Zero-shot transfer=true2025.03 | 43.7 | 46.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MASA-R50Backbone=ResNet-50, Zero-shot transfer=true2025.03 | 42.7 | 45.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AEDDetector=RegionCLIP2024.09 | 38.1 | 37 | 56.7 | 16.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETerarchitecture=HTC2022.07 | 37.53 | 36.85 | 57.53 | 15.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETerbackbone=SwinT2022.07 | 36.71 | 34.61 | 52.1 | 15.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| OVTrackPublication=CVPR20232024.09 | 36.7 | 34.7 | 49.3 | 18.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| OVTrackBackbone=ResNet50, Detector=Faster R-CNN, Video-based training=false2023.04 | 36.7 | 34.7 | 49.3 | 18.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UNINEXTPublication=CVPR2023, Backbone=ResNet-502024.09 | 35.5 | 31.9 | 43.3 | 17.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETer2022.07 | 35.02 | 33.25 | 51.58 | 13.16 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETerPublication=ECCV20222024.09 | 35 | 33.3 | 51.6 | 13.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETerBackbone=ResNet101, Detector=Faster R-CNN, Video-based training=true2023.04 | 35 | 33.3 | 51.6 | 13.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AOABackbone=Other, Detector=Faster R-CNN, Video-based training=true2023.04 | 30.6 | 25.3 | 23.4 | 21.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AOA2022.07 | 30.56 | 25.27 | 23.4 | 21.86 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| QDTrackarchitecture=HTC2022.07 | 27.43 | 32.79 | 56.21 | 14.73 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| QDTrackPublication=CVPR20212024.09 | 27.4 | 30 | 50.5 | 12.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| QDTrackBackbone=ResNet101, Detector=Faster R-CNN, Video-based training=true2023.04 | 27.4 | 30 | 50.5 | 12.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| QDTrack2022.07 | 27.36 | 30 | 50.53 | 12.11 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| QDTrackbackbone=SwinT2022.07 | 27.27 | 31.22 | 51.32 | 15.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Tracktor++2022.07 | 22.81 | 27.97 | 49.04 | 12.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Tracktor++Publication=ECCV20222024.09 | 22.8 | 28 | 49 | 12.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Tracktor++Backbone=ResNet101, Detector=Faster R-CNN, Video-based training=true2023.04 | 22.8 | 28 | 49 | 12.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DeepSORT2022.07 | 17.52 | 25.98 | 48.35 | 12.09 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DeepSORTPublication=ICIP20172024.09 | 17.5 | 26 | 48.4 | 12.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DeepSORTBackbone=ResNet101, Detector=Faster R-CNN, Video-based training=true2023.04 | 17.5 | 26 | 48.4 | 12.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SORT2022.07 | 14.32 | 24.84 | 48.13 | 12.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SORT-TAOBackbone=ResNet101, Detector=Faster R-CNN, Video-based training=true2023.04 | 14.3 | 24.8 | 48.1 | 12.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TracktorPublication=ICCV20192024.09 | 13 | 24.2 | 47.4 | 12.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TracktorBackbone=ResNet101, Detector=Faster R-CNN, Video-based training=true2023.04 | 13 | 24.2 | 47.4 | 12.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Tracktor2022.07 | 12.96 | 24.15 | 47.41 | 12.08 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AEDDetector=RegionCLIP [12], Publication=Ours2024.09 | — | — | — | — | 37.3 | 56.4 | 38.4 | 17 | 34.4 | 55.9 | 38.5 | 870 | — | — | — | — | — | — | — | — | |
| AEDDetector=Co-DETR [37], Publication=Ours2024.09 | — | — | — | — | 55.7 | 71.5 | 52.2 | 43.5 | 51.4 | 74.4 | 51.8 | 2,810 | — | — | — | — | — | — | — | — | |
| AOAFPS=12022.03 | — | — | — | — | — | — | — | — | — | — | — | — | 25.8 | — | — | — | — | — | — | — | |
| BIVSupervision=Fully-supervised, Evaluation Protocol=In-domain, Tracking Paradigm=Online2024.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 21.6 | 10.4 | 16.1 | |
| ByteTrackDetector=RegionCLIP [12], Publication=ECCV20222024.09 | — | — | — | — | 29.4 | 52.3 | 19.8 | 16 | 26.5 | 50.8 | 20.9 | 800 | — | — | — | — | — | — | — | — | |
| ByteTrackDetector=Co-DETR [37], Publication=ECCV20222024.09 | — | — | — | — | 41 | 62.3 | 15.9 | 44.8 | 37.2 | 65.9 | 17 | 2,870 | — | — | — | — | — | — | — | — | |
| ByteTrackBackbone=ResNet-50, Detector=Same detector, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 28.2 | 50.4 | 18.1 | 16 | 22 | 48.2 | 16.6 | 1 | — | — | — | — | — | — | — | — | |
| COVTrackBackbone=ResNet-50, Detector=Same detector, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 39.6 | 57.3 | 42 | 19.6 | 34.3 | 58.2 | 41.3 | 3.5 | — | — | — | — | — | — | — | — | |
| COVTrack++Backbone=ResNet-50, Detector=Same detector, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 40.3 | 58.6 | 42.6 | 19.6 | 35.4 | 60.1 | 42.6 | 3.5 | — | — | — | — | — | — | — | — | |
| DeepSORTDetector=RegionCLIP [12], Publication=ICIP20172024.09 | — | — | — | — | 28.4 | 52.5 | 15.6 | 17 | 24.5 | 49.2 | 15.3 | 900 | — | — | — | — | — | — | — | — | |
| DeepSORTDetector=Co-DETR [37], Publication=ICIP20172024.09 | — | — | — | — | 45.8 | 70.5 | 28.1 | 38.9 | 42.2 | 72.7 | 29 | 2,470 | — | — | — | — | — | — | — | — | |
| DeepSORT (ViLD)Backbone=ResNet-50, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 26.9 | 47.1 | 15.8 | 17.7 | 21.1 | 46.4 | 14.7 | 2.3 | — | — | — | — | — | — | — | — | |
| GTRSupervision=Fully-supervised, Evaluation Protocol=In-domain, Tracking Paradigm=Offline, Detection Source=Original2024.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 22.5 | — | — | |
| MASA (R50)Backbone=ResNet-50, Detector=Same detector, Classes (Novel)=False, Classes (Base)=False2026.03 | — | — | — | — | 36.9 | 55.1 | 36.4 | 19.3 | 30 | 54.2 | 34.6 | 1 | — | — | — | — | — | — | — | — | |
| NetTrackDetector=GroundingDINO [13], Publication=CVPR20242024.09 | — | — | — | — | 33 | 45.7 | 28.6 | 24.8 | 32.6 | 51.3 | 33 | 1,330 | — | — | — | — | — | — | — | — | |
| OC-SORTBackbone=ResNet-50, Detector=Same detector, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 28.9 | 51.4 | 19.8 | 15.4 | 23.7 | 49.6 | 20.4 | 1.1 | — | — | — | — | — | — | — | — | |
| OursBackbone=ResNet-1012022.06 | — | — | — | — | — | — | — | — | — | — | — | — | 19.9 | — | — | 8.3 | 9.6 | — | — | — | |
| Ours-DeticSupervision=Self-supervised, Evaluation Protocol=Zero-shot, Detection Source=Same detections as GTR, Tracking Paradigm=Online2024.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 22 | 12.2 | 17.1 | |
| Ours-DeticSupervision=Self-supervised, Evaluation Protocol=Zero-shot, Detection Source=Original detections, Tracking Paradigm=Online2024.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 30.9 | 18 | 24.4 | |
| Ours-Grounding-DINOSupervision=Self-supervised, Evaluation Protocol=Zero-shot, Detection Source=Same detections as GTR, Tracking Paradigm=Online2024.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 22.8 | 12.3 | 17.6 | |
| Ours-SAM-BSupervision=Self-supervised, Evaluation Protocol=Zero-shot, Detection Source=Same detections as GTR, Tracking Paradigm=Online2024.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 23.9 | 13 | 18.4 | |
| Ours-SAM-HSupervision=Self-supervised, Evaluation Protocol=Zero-shot, Detection Source=Same detections as GTR, Tracking Paradigm=Online2024.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 22.9 | 12.1 | 17.5 | |
| OVSORTBackbone=ResNet-50, Detector=Same detector, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 38.2 | 55.3 | 39.9 | 19.4 | 30.8 | 53 | 37.6 | 1.9 | — | — | — | — | — | — | — | — | |
| OVTRBackbone=ResNet-50, Detector=Same detector, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 36.6 | 52.2 | 37.6 | 20.1 | 31.4 | 54.4 | 34.5 | 5.4 | — | — | — | — | — | — | — | — | |
| OVTrackDetector=Integrated, Publication=CVPR20232024.09 | — | — | — | — | 35.5 | 49.3 | 36.9 | 20.2 | 27.8 | 48.8 | 33.6 | 150 | — | — | — | — | — | — | — | — | |
| OVTrackDetector=RegionCLIP [12], Publication=CVPR20232024.09 | — | — | — | — | 36.3 | 53.9 | 36.3 | 18.7 | 32 | 51.4 | 33.2 | 1,140 | — | — | — | — | — | — | — | — | |
| OVTrackDetector=Co-DETR [37], Publication=CVPR20232024.09 | — | — | — | — | 54.5 | 70.3 | 47.5 | 45.6 | 49.1 | 73.7 | 46.3 | 2,720 | — | — | — | — | — | — | — | — | |
| OVTrackBackbone=ResNet-50, Detector=Same detector, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 35.5 | 49.3 | 36.9 | 20.2 | 27.8 | 48.8 | 33.6 | 1.5 | — | — | — | — | — | — | — | — | |
| QDTrackDetector=Integrated, Publication=CVPR20212024.09 | — | — | — | — | 27.1 | 45.6 | 24.7 | 11 | 22.5 | 42.7 | 24.4 | 40 | — | — | — | — | — | — | — | — | |
| QDTrackBackbone=ResNet-1012022.06 | — | — | — | — | — | — | — | — | — | — | — | — | 15.8 | — | — | 6.4 | 7.3 | — | — | — | |
| QDTrackSupervision=Fully-supervised, Evaluation Protocol=In-domain, Tracking Paradigm=Online2024.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 15.9 | 5 | 10.6 | |
| QDTrackBackbone=ResNet-50, Classes (Novel)=True, Classes (Base)=True2026.03 | — | — | — | — | 27.1 | 45.6 | 24.7 | 11 | 22.5 | 42.7 | 24.4 | 0.4 | — | — | — | — | — | — | — | — | |
| SLAckBackbone=ResNet-50, Detector=Same detector, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 37.2 | 55 | 37.6 | 19.1 | 31.1 | 54.3 | 37.8 | 1.3 | — | — | — | — | — | — | — | — | |
| SORTBackbone=ResNet-1012022.06 | — | — | — | — | — | — | — | — | — | — | — | — | 13.2 | — | — | — | — | — | — | — | |
| SORT_TAOFPS=15.22022.03 | — | — | — | — | — | — | — | — | — | — | — | — | 13.2 | — | — | — | — | — | — | — | |
| SORT-TAOSupervision=Fully-supervised, Evaluation Protocol=In-domain, Tracking Paradigm=Online2024.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 13.2 | — | — | |
| TACSupervision=Fully-supervised, Evaluation Protocol=In-domain, Tracking Paradigm=Online2024.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 17.7 | 5.8 | 7.3 | |
| TETerDetector=Integrated, Publication=ECCV20222024.09 | — | — | — | — | 30.3 | 47.4 | 31.6 | 12.1 | 25.7 | 45.9 | 31.1 | 20 | — | — | — | — | — | — | — | — | |
| TETerBackbone=ResNet-50, Classes (Novel)=True, Classes (Base)=True2026.03 | — | — | — | — | 30.3 | 47.4 | 31.6 | 12.1 | 25.7 | 45.9 | 31.1 | 0.2 | — | — | — | — | — | — | — | — | |
| Tracktor++Detector=RegionCLIP [12], Publication=ICCV20192024.09 | — | — | — | — | 29.6 | 52.4 | 19.6 | 16.9 | 25.7 | 50.1 | 18.9 | 810 | — | — | — | — | — | — | — | — | |
| Tracktor++ (ViLD)Backbone=ResNet-50, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 28.3 | 47.4 | 20.5 | 17 | 22.7 | 46.7 | 19.3 | 2.2 | — | — | — | — | — | — | — | — | |
| TRACTBackbone=ResNet-50, Detector=Same detector, Classes (Novel)=False, Classes (Base)=True2026.03 | — | — | — | — | 38.5 | 55 | 39 | 21.5 | 31.3 | 52.7 | 37.8 | 3.4 | — | — | — | — | — | — | — | — |