Multi-Object Tracking on BDD100K (val)
58.8mIDF1SUSHI
Evaluation Results
| Method | Links | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SUSHI2022.12 | 58.8 | 45.8 | 68.4 | 75.6 | — | — | 8,556 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UNINEXT-HSplit=val, Backbone=ViT-H2023.03 | 56.7 | 44.2 | 67.1 | 69.9 | — | — | 10,222 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UNINEXT-Hsupervision_type=Fully-supervised, evaluation_protocol=In-domain, provided_detections=true2024.06 | 56.7 | 44.2 | — | 69.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MOTRv2YOLOX proposals=true2022.11 | 56.5 | 43.6 | 65.6 | 72.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MOTRv2Venue=CVPR 232024.06 | 56.5 | 43.6 | 65.6 | — | — | — | — | — | — | — | — | — | — | — | — | 72.7 | — | — | — | — | |
| MOTRv2Backbone=YOLOX-X+ResNet-50, Pre-train=BDD100K+COCO, Objectness threshold=0.42023.11 | 56.5 | 43.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GeneralTrack2024.06 | 56.2 | 46.4 | 68.8 | 72.7 | — | — | 8,496 | 11,830 | 2,035 | — | — | — | — | — | 46.9 | 63.1 | — | — | — | — | |
| Ours-Deticsupervision_type=Self-supervised, evaluation_protocol=Zero-shot, provided_detections=true2024.06 | 55.8 | 44.6 | — | 71.3 | — | — | — | — | — | 54.4 | — | 52.9 | — | — | — | — | — | — | — | — | |
| Ours-Deticsupervision=Self-supervised, protocol=zero-shot, backbone=Swin-B2024.06 | 55.8 | 44.6 | — | 71.3 | — | — | — | — | — | 54.4 | — | 52.9 | — | — | — | — | — | — | — | — | |
| GHOST2022.06 | 55.6 | 44.9 | 68.1 | 70.9 | — | — | — | — | — | — | — | — | — | — | 45.7 | 61.7 | — | — | — | — | |
| GHOSTVenue=CVPR 232024.06 | 55.6 | 44.9 | 68.1 | 70.9 | — | — | — | — | — | — | — | — | — | — | 45.7 | 61.7 | — | — | — | — | |
| Ours-Grounding-DINOsupervision_type=Self-supervised, evaluation_protocol=Zero-shot, provided_detections=true2024.06 | 55.6 | 44.5 | — | 71.7 | — | — | — | — | — | 54.5 | — | 52.7 | — | — | — | — | — | — | — | — | |
| Ours-SAM-Bsupervision_type=Self-supervised, evaluation_protocol=Zero-shot, provided_detections=true2024.06 | 55.6 | 44.1 | — | 71.6 | — | — | — | — | — | 54 | — | 52.6 | — | — | — | — | — | — | — | — | |
| Ours-Grounding-DINOsupervision=Self-supervised, protocol=zero-shot, backbone=Swin-B2024.06 | 55.6 | 44.5 | — | 71.7 | — | — | — | — | — | 54.5 | — | 52.7 | — | — | — | — | — | — | — | — | |
| Ours-SAM-B+supervision=Self-supervised, protocol=zero-shot, backbone=SAM-ViT-B2024.06 | 55.6 | 44.1 | — | 71.6 | — | — | — | — | — | 54 | — | 52.6 | — | — | — | — | — | — | — | — | |
| Ours-SAM-Hsupervision_type=Self-supervised, evaluation_protocol=Zero-shot, provided_detections=true2024.06 | 55.3 | 44.5 | — | 71.7 | — | — | — | — | — | 54.2 | — | 51.9 | — | — | — | — | — | — | — | — | |
| Ours-SAM-Hsupervision=Self-supervised, protocol=zero-shot, backbone=SAM-ViT-H2024.06 | 55.3 | 44.5 | — | 71.7 | — | — | — | — | — | 54.2 | — | 51.9 | — | — | — | — | — | — | — | — | |
| UNINEXT-LSplit=val, Backbone=ConvNeXt-L2023.03 | 54.9 | 41.8 | 64.6 | 68.7 | — | — | 9,134 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ByteTrack2021.10 | 54.8 | 45.5 | 69.1 | 70.4 | 92,805 | 34,998 | 9,140 | 9,626 | 3,005 | — | — | — | — | — | — | — | — | — | — | — | |
| ByteTrack2022.12 | 54.8 | 45.5 | 69.1 | 70.4 | — | — | 9,140 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ByteTrackDetector=YOLOX-X, Base model=Modified CSP, Pre-trained weights=COCO2022.10 | 54.8 | 45.5 | 69.1 | 70.4 | 92,805 | 34,998 | 9,140 | 9,626 | 3,005 | — | — | — | — | — | — | — | — | — | — | — | |
| ByteTrackVenue=ECCV'222024.06 | 54.8 | 45.2 | 69.1 | 70.4 | — | — | 9,140 | 9,626 | 3,005 | — | — | — | — | — | 45.3 | 61.3 | — | — | — | — | |
| ByteTracksupervision_type=Fully-supervised, evaluation_protocol=In-domain, provided_detections=true2024.06 | 54.8 | 45.5 | — | 70.4 | — | — | — | — | — | 55.7 | — | 51.5 | — | — | — | — | — | — | — | — | |
| ByteTracksupervision=Fully-supervised, protocol=in-domain2024.06 | 54.8 | 45.5 | — | 70.4 | — | — | — | — | — | 55.7 | — | 51.5 | — | — | — | — | — | — | — | — | |
| Ours-R50+supervision=Self-supervised, protocol=zero-shot, backbone=ResNet-502024.06 | 54.8 | 44.2 | — | 71.3 | — | — | — | — | — | 54 | — | 51.7 | — | — | — | — | — | — | — | — | |
| ByteTrack2022.06 | 54.6 | 45.2 | 68.7 | 70.2 | — | — | — | — | — | — | — | — | — | — | 45.4 | 61.6 | — | — | — | — | |
| QDTrackDetector=YOLOX-X, Base model=Modified CSP, Pre-trained weights=COCO2022.10 | 54.3 | 42.1 | 68.2 | 73.3 | 83,395 | 48,798 | 8,478 | 10,925 | 2,272 | — | — | — | — | — | — | — | — | — | — | — | |
| UnicornSplit=val2022.07 | 54 | 41.2 | 66.6 | 71.3 | 95,454 | 41,648 | 10,876 | 10,296 | 2,505 | — | — | — | — | 41.4 | — | — | — | — | — | — | |
| Unicorn2022.12 | 54 | 41.2 | 66.6 | 71.3 | — | — | 10,876 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Unicorn2022.11 | 54 | 41.2 | 66.6 | 71.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UnicornSplit=val2023.03 | 54 | 41.2 | 66.6 | 71.3 | — | — | 10,876 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UnicornVenue=ECCV'222024.06 | 54 | 41.2 | 66.6 | — | — | — | 10,876 | 10,296 | 2,505 | — | — | — | — | — | — | 71.3 | — | — | — | — | |
| MOTRsupervision_type=Fully-supervised, evaluation_protocol=In-domain, provided_detections=false2024.06 | 54 | 32.3 | — | 65.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Unicornsupervision_type=Fully-supervised, evaluation_protocol=In-domain, provided_detections=false2024.06 | 54 | 41.2 | — | 71.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETer2022.07 | 53.3 | 39.1 | — | — | — | — | — | — | — | 50.83 | 47.16 | 52.89 | 52.44 | — | — | — | — | — | — | — | |
| TETer2022.12 | 53.3 | 39.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETer2022.11 | 53.3 | 39.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETerDetector=FRCNN, Base model=R50-FPN2022.10 | 53.3 | 39.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETerVenue=ECCV'222024.06 | 53.3 | 39.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETersupervision_type=Fully-supervised, evaluation_protocol=In-domain, provided_detections=false2024.06 | 53.3 | 39.1 | — | 71.1 | — | — | — | — | — | 50.8 | — | 52.9 | — | — | — | — | — | — | — | — | |
| QDTrackDetector=FRCNN, Base model=R50-FPN2022.10 | 52.9 | 37.7 | 65.7 | 72.7 | 104,861 | 41,355 | 5,640 | 9,649 | 2,874 | — | — | — | — | — | — | — | — | — | — | — | |
| ContrasTRBackbone=Swin-L, Pre-train=BDD100K, Objectness threshold=0.42023.11 | 52.9 | 41.7 | — | — | — | — | 6,363 | — | — | — | — | — | — | — | 44.4 | — | — | — | — | — | |
| QDTrack2022.07 | 51.6 | 36.6 | — | — | — | — | — | — | — | 47.84 | 45.86 | 48.47 | 49.2 | — | — | — | — | — | — | — | |
| QDTrack2022.06 | 51.5 | 36.3 | 63.7 | 71.4 | — | — | — | — | — | — | — | — | — | — | 41.7 | 60.9 | — | — | — | — | |
| QDTrack2021.10 | 50.8 | 36.6 | 63.5 | 71.5 | 108,614 | 46,621 | 6,262 | 9,481 | 3,034 | — | — | — | — | — | — | — | — | — | — | — | |
| QDTrackSplit=val2022.07 | 50.8 | 36.6 | 63.5 | 71.5 | 108,614 | 46,621 | 6,262 | 9,481 | 3,034 | — | — | — | — | 32.6 | — | — | — | — | — | — | |
| QDTrack2022.12 | 50.8 | 36.6 | 63.5 | 71.5 | — | — | 6,262 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| QDTrack2022.11 | 50.8 | 36.6 | 63.5 | 71.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| QDTrackSplit=val2023.03 | 50.8 | 36.6 | 63.5 | 71.5 | — | — | 6,262 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| QDTrackVenue=CVPR'212024.06 | 50.8 | 36.6 | 63.5 | — | — | — | 6,262 | 9,481 | 3,034 | — | — | — | — | — | — | 71.5 | — | — | — | — | |
| QDTracksupervision_type=Fully-supervised, evaluation_protocol=In-domain, provided_detections=false2024.06 | 50.8 | 36.6 | — | 71.5 | — | — | — | — | — | 47.8 | — | 48.5 | — | — | — | — | — | — | — | — | |
| QDTrackSplit=val2020.06 | 50.8 | 36.6 | 63.5 | 71.5 | 108,614 | 46,621 | 6,262 | 9,481 | 3,034 | — | — | — | — | 32.6 | — | — | — | — | — | — | |
| DeepSORT2022.07 | 49.3 | 35.2 | — | — | — | — | — | — | — | 48.03 | 46.36 | 46.69 | 51.04 | — | — | — | — | — | — | — | |
| DeepSORTDetector=FRCNN, Base model=R50-FPN2022.10 | 49.3 | 35.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ContrasTRBackbone=ResNet-50, Pre-train=BDD100K, Objectness threshold=0.42023.11 | 49.2 | 36.7 | — | — | — | — | 6,695 | — | — | — | — | — | — | — | 40.8 | — | — | — | — | — | |
| MOTR*YOLOX proposals=false2022.11 | 48.2 | 35.5 | 59.6 | 68.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MOTRv2Backbone=ResNet-50, Pre-train=COCO, Objectness threshold=0.42023.11 | 48.2 | 35.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ContrasTR w/o CPTBackbone=ResNet-50, Pre-train=COCO, Objectness threshold=0.42023.11 | 48.1 | 36.4 | — | — | — | — | 6,067 | — | — | — | — | — | — | — | 40.2 | — | — | — | — | — | |
| MOTR2022.11 | 44.8 | 32.3 | 56.2 | 65.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MOTRVenue=ECCV'222024.06 | 44.8 | 32.3 | 56.2 | — | — | — | — | — | — | — | — | — | — | — | — | 65.8 | — | — | — | — | |
| Yu et al.2021.10 | 44.5 | 25.9 | 56.9 | 66.8 | 122,406 | 52,372 | 8,315 | 8,396 | 3,795 | — | — | — | — | — | — | — | — | — | — | — | |
| Yu et al.Split=val2022.07 | 44.5 | 25.9 | 56.9 | 66.8 | 122,406 | 52,372 | 8,315 | 8,396 | 3,795 | — | — | — | — | 28.1 | — | — | — | — | — | — | |
| Yu et al.2022.12 | 44.5 | 25.9 | 56.9 | 66.8 | — | — | 8,315 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Yu et al.2022.11 | 44.5 | 25.9 | 56.9 | 66.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Yu et al.Split=val2023.03 | 44.5 | 25.9 | 56.9 | 66.8 | — | — | 8,315 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Yu et alDetector=FRCNN, Base model=DLA-342022.10 | 44.5 | 25.9 | 56.9 | 66.8 | 122,406 | 52,372 | 8,315 | 8,396 | 3,795 | — | — | — | — | — | — | — | — | — | — | — | |
| Yu et al.Split=val2020.06 | 44.5 | 25.9 | 56.9 | 66.8 | 122,406 | 52,372 | 8,315 | 8,396 | 3,795 | — | — | — | — | 28.1 | — | — | — | — | — | — | |
| MOTR2022.12 | 43.5 | 32 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MOTRBackbone=ResNet-50, Pre-train=COCO, Objectness threshold=0.42023.11 | 43.5 | 32 | — | — | — | — | 3,493 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Yu et al.2022.06 | 25.9 | — | — | 56.9 | — | — | — | — | — | — | — | — | — | — | 44.5 | 66.8 | — | — | — | — | |
| CO-MOT2023.05 | — | — | — | — | — | — | — | — | — | 52.8 | 38.7 | 56.2 | 63.6 | — | — | — | — | — | — | — | |
| CO-MOT2023.05 | — | — | — | — | — | — | — | — | — | 52.8 | 38.7 | 56.2 | 63.6 | — | — | — | — | — | — | — | |
| COVTrackinference_mode=zero-shot, training_dataset=TAO2026.03 | — | — | — | — | — | — | — | — | — | 45.8 | 44.1 | 39.4 | 53.9 | — | — | — | — | — | — | — | |
| COVTrack++inference_mode=zero-shot, training_dataset=TAO2026.03 | — | — | — | — | — | — | — | — | — | 46.7 | 45.7 | 40.7 | 53.8 | — | — | — | — | — | — | — | |
| DeepSORT2023.05 | — | — | — | — | — | — | — | — | — | 48 | 46.4 | 46.7 | 51 | — | — | — | — | — | — | — | |
| DeepSORT2023.05 | — | — | — | — | — | — | — | — | — | 48 | 46.4 | 46.7 | 51 | — | — | — | — | — | — | — | |
| DeepSORT2023.07 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 48 | 46.4 | 46.7 | — | |
| MASAPublication=CVPR2024, Classification Setting=8 classes2025.04 | — | — | — | 71.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MeMOTR2023.07 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 40.4 | — | 53.6 | 38.1 | 56.7 | 52.6 | |
| MOTR2022.06 | — | — | 32 | — | — | — | — | — | — | — | — | — | — | — | — | 43.5 | — | — | — | — | |
| MOTR2023.07 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 37 | — | 50.7 | 35.8 | 51 | 47.3 | |
| MOTRv22023.05 | — | — | — | — | — | — | — | — | — | 54.9 | 49.5 | 51.9 | 63.1 | — | — | — | — | — | — | — | |
| MOTRv22023.05 | — | — | — | — | — | — | — | — | — | 54.9 | 49.5 | 51.9 | 63.1 | — | — | — | — | — | — | — | |
| MOTRv2Publication=CVPR2022, Classification Setting=8 classes2025.04 | — | — | 65.6 | 72.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| OVTRinference_mode=zero-shot, training_dataset=TAO2026.03 | — | — | — | — | — | — | — | — | — | 43.1 | 42 | 37.1 | 50.1 | — | — | — | — | — | — | — | |
| OVTrackTraining Data=LVIS, Logit masking=false2023.04 | — | — | — | — | — | — | — | — | — | 42.5 | 41 | 36.7 | 49.7 | — | — | — | — | — | — | — | |
| OVTrackinference_mode=zero-shot, training_dataset=TAO2026.03 | — | — | — | — | — | — | — | — | — | 42.5 | 41 | 36.7 | 49.7 | — | — | — | — | — | — | — | |
| QDtrackPublication=TPAMI2023, Classification Setting=8 classes2025.04 | — | — | 63.5 | 71.5 | — | — | 6,262 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| QDTrackTraining Data=LVIS, TAO, Logit masking=true2023.04 | — | — | — | — | — | — | — | — | — | 35.6 | 38.1 | 28.5 | 40.2 | — | — | — | — | — | — | — | |
| QDTrackTraining Data=LVIS, TAO, Logit masking=false2023.04 | — | — | — | — | — | — | — | — | — | 32 | 25.9 | 27.8 | 42.4 | — | — | — | — | — | — | — | |
| QDTrack2023.05 | — | — | — | — | — | — | — | — | — | 47.8 | 45.8 | 48.5 | 49.2 | — | — | — | — | — | — | — | |
| QDTrack2023.05 | — | — | — | — | — | — | — | — | — | 47.8 | 45.8 | 48.5 | 49.2 | — | — | — | — | — | — | — | |
| QDTrack2023.07 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 47.8 | 45.9 | 48.5 | — | |
| QDTrackinference_mode=zero-shot, training_dataset=TAO2026.03 | — | — | — | — | — | — | — | — | — | 32 | 25.9 | 27.8 | 42.4 | — | — | — | — | — | — | — | |
| SAM2MOTClassification Setting=8 classes, Backbone=Co-DINO-L, Fine-tuning status=no fine-tuning2025.04 | — | — | 57.5 | 70.8 | — | — | 5,587 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SAM2MOTClassification Setting=8 classes, Backbone=Grounding-DINO-L, Fine-tuning status=no fine-tuning2025.04 | — | — | 44.1 | 63.6 | — | — | 5,184 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SAM2MOTClassification Setting=3 classes, Backbone=Co-DINO-L, Fine-tuning status=no fine-tuning2025.04 | — | — | 63 | 73.7 | — | — | 5,755 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SAM2MOTClassification Setting=3 classes, Backbone=Grounding-DINO-L, Fine-tuning status=no fine-tuning2025.04 | — | — | 58 | 71 | — | — | 5,464 | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TETerTraining Data=LVIS, TAO, Logit masking=true2023.04 | — | — | — | — | — | — | — | — | — | 36.1 | 36.4 | 31.9 | 40.2 | — | — | — | — | — | — | — | |
| TETerTraining Data=LVIS, TAO, Logit masking=false2023.04 | — | — | — | — | — | — | — | — | — | 33.2 | 24.5 | 31.8 | 43.4 | — | — | — | — | — | — | — |