Multiview Pedestrian Detection on Wildtrack (test)
94.2MODAMVDGC
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MVDGCExtra Info=false2026.06 | 94.2 | 83.8 | 97.4 | 96.9 | 97.1 | |
| MVFPExtra Info=false2026.06 | 94.1 | 78.8 | 96.4 | 97.7 | 97 | |
| MVTTExtra Info=true2026.06 | 94.1 | 81.3 | 97.6 | 96.5 | 97 | |
| PVH-EncExtra Info=true2026.06 | 93.6 | 82.4 | 96.6 | 97 | 96.8 | |
| 3DROMEvaluation protocol=single-scene, Trained on=Wildtrack2024.05 | 93.5 | 75.9 | 97.2 | 96.2 | 96.7 | |
| 3DROMExtra Info=false2026.06 | 93.5 | 75.9 | 97.2 | 96.2 | 96.7 | |
| MVAugExtra Info=false2026.06 | 93.2 | 79.8 | 96.3 | 97 | 96.6 | |
| TrackTacularExtra Info=true2026.06 | 93.2 | 77.5 | 97.3 | 95.8 | 96.5 | |
| MVFlow2022.10 | 91.9 | — | 96.4 | 95.7 | — | |
| MVDeTrImageNet (pre-train)=true2021.09 | 91.5 | 82.1 | 97.4 | 94 | — | |
| MVDeTr2022.10 | 91.5 | — | 97.4 | 94 | — | |
| MVDeTrEvaluation protocol=single-scene, Trained on=Wildtrack2024.05 | 91.5 | 82.1 | 97.4 | 94 | 95.7 | |
| MVDeTr (ours)2021.08 | 91.5 | 82.1 | 97.4 | 94 | — | |
| MVDetrExtra Info=false2026.06 | 91.5 | 82.1 | 97.4 | 94 | 95.7 | |
| EarlybirdExtra Info=false2026.06 | 91.2 | 81.8 | 94.9 | 96.3 | 95.6 | |
| SHOTImageNet (pre-train)=false2021.09 | 90.2 | 76.5 | 96.1 | 94 | — | |
| SHOT2022.10 | 90.2 | — | 96.1 | 94 | — | |
| SHOTEvaluation protocol=single-scene, Trained on=Wildtrack2024.05 | 90.2 | 76.5 | 96.1 | 94 | 95 | |
| MVDeTrAggregation=convolution2021.08 | 90.2 | 81.6 | 96.9 | 93.2 | — | |
| SHOTExtra Info=false2026.06 | 90.2 | 76.5 | 96.1 | 94 | 95 | |
| MVDeTrPer-view loss=false2021.08 | 89.9 | 81.8 | 97.7 | 92.1 | — | |
| MVDeTrAugmentation=false2021.08 | 89.5 | 81.2 | 94 | 95.6 | — | |
| MVDetAugmentation=true2021.08 | 89 | 75.5 | 93.5 | 95.5 | — | |
| VolumetricEvaluation protocol=single-scene, Trained on=Wildtrack2024.05 | 88.6 | 73.8 | 95.3 | 93.2 | 94.2 | |
| MVDetMultiview aggregation=feature maps, Spatial aggregation=large kernel convolution2020.07 | 88.2 | 75.7 | 94.7 | 93.6 | — | |
| MVDetImageNet (pre-train)=false2021.09 | 88.2 | 75.7 | 94.7 | 93.6 | — | |
| MVDet2022.10 | 88.2 | — | 94.7 | 93.6 | — | |
| MVDetEvaluation protocol=single-scene, Trained on=Wildtrack2024.05 | 88.2 | 75.7 | 94.7 | 93.6 | 94.1 | |
| MVDet2021.08 | 88.2 | 75.7 | 94.7 | 93.6 | — | |
| MVDetExtra Info=false2026.06 | 88.2 | 75.7 | 94.7 | 93.6 | 94.1 | |
| Proposed MethodImageNet (pre-train)=false2021.09 | 87.2 | 74.5 | 93.8 | 93.4 | — | |
| Proposed Method (DropView)ImageNet (pre-train)=true2021.09 | 86.7 | 76.2 | 95.1 | 91.4 | — | |
| Proposed MethodImageNet (pre-train)=true2021.09 | 85.4 | 76.7 | 95.2 | 89.9 | — | |
| Supervised View-Wise Contribution WeightingEvaluation protocol=cross-scene, Trained on=CVCS, Finetuning=5% of labeled data, Domain adaptation=true (ft+da)2024.05 | 78.9 | 73.6 | 88.7 | 90.4 | 89.5 | |
| MVDet (w/o large kernel)Multiview aggregation=feature maps, Spatial aggregation=N/A2020.07 | 76.9 | 71.6 | 84.5 | 93.5 | — | |
| Deep-OcclusionMultiview aggregation=anchor box features, Spatial aggregation=CRF + mean-field inference2020.07 | 74.1 | 53.8 | 95 | 80 | — | |
| Deep-OcclusionImageNet (pre-train)=false2021.09 | 74.1 | 53.8 | 95 | 80 | — | |
| DeepOcclusion2022.10 | 74.1 | — | 95 | 80 | — | |
| DeepOcc.Evaluation protocol=single-scene, Trained on=Wildtrack2024.05 | 74.1 | 53.8 | 95 | 80 | 86.9 | |
| Deep-Occlusion2021.08 | 74.1 | 53.8 | 95 | 80 | — | |
| Deep-OccExtra Info=false2026.06 | 74.1 | 53.8 | 95 | 80 | 86.8 | |
| Supervised View-Wise Contribution WeightingEvaluation protocol=cross-scene, Trained on=CVCS, Finetuning=5% of labeled data (ft)2024.05 | 73.9 | 72.4 | 86.8 | 87.2 | 87 | |
| MVDet (project results)Multiview aggregation=detection results, Spatial aggregation=large kernel convolution2020.07 | 68.2 | 71.9 | 85.9 | 81.2 | — | |
| DeepMCDMultiview aggregation=anchor box features, Spatial aggregation=N/A2020.07 | 67.8 | 64.2 | 85 | 82 | — | |
| DeepMCDImageNet (pre-train)=false2021.09 | 67.8 | 64.2 | 85 | 82 | — | |
| DeepMCDEvaluation protocol=single-scene, Trained on=Wildtrack2024.05 | 67.8 | 64.2 | 85 | 82 | 83.5 | |
| DeepMCD2021.08 | 67.8 | 64.2 | 85 | 82 | — | |
| DeepMCDExtra Info=false2026.06 | 67.8 | 64.2 | 85 | 82 | 83.5 | |
| Lima et al.ImageNet (pre-train)=false2021.09 | 56.9 | 67.3 | 80.8 | 74.6 | — | |
| Lopez-Cifuentes et al.ImageNet (pre-train)=false2021.09 | 39 | 55 | — | — | — | |
| MVDet (project images)Multiview aggregation=RGB image pixels, Spatial aggregation=large kernel convolution2020.07 | 26.8 | 45.6 | 84.2 | 33 | — | |
| POM-CNNMultiview aggregation=detection results, Spatial aggregation=mean-field inference2020.07 | 23.2 | 30.5 | 75 | 55 | — | |
| POM-CNNImageNet (pre-train)=false2021.09 | 23.2 | 30.5 | 75 | 55 | — | |
| POM-CNNEvaluation protocol=single-scene, Trained on=Wildtrack2024.05 | 23.2 | 30.5 | 75 | 55 | 63.5 | |
| POM-CNN2021.08 | 23.2 | 30.5 | 75 | 55 | — | |
| POM-CNNExtra Info=false2026.06 | 23.2 | 30.5 | 75 | 55 | 63.5 | |
| RCNN & clusteringMultiview aggregation=detection results, Spatial aggregation=clustering2020.07 | 11.3 | 18.4 | 68 | 43 | — | |
| RCNN ClusteringImageNet (pre-train)=false2021.09 | 11.3 | 18.4 | 68 | 43 | — | |
| RCNNEvaluation protocol=single-scene, Trained on=Wildtrack2024.05 | 11.3 | 18.4 | 68 | 43 | 52.7 | |
| RCNN & clustering2021.08 | 11.3 | 18.4 | 68 | 43 | — | |
| RCNN-basedExtra Info=false2026.06 | 11.3 | 18.4 | 68 | 43 | 52.7 |