Video Object Detection on ImageNet VID v1.0 (val)
93.2AP50YOLOV++
Evaluation Results
| Method | Links | |
|---|---|---|
| YOLOV++Backbone=Modified CSP v5, Base detector=YOLOX-FocalL, Post-processing=REPP [46]2024.07 | 93.2 | |
| YOLOV++Backbone=FocalLarge, Base detector=YOLOX-FocalL, Time (ms)=27.6, Post-processing=None2024.07 | 92.9 | |
| ObjectsBackbone=SwinBase, Base detector=Deformable DETR, Post-processing=None2024.07 | 91.3 | |
| YOLOV++Backbone=Modified CSP v5, Base detector=YOLOX-SwinB, Post-processing=REPP [46]2024.07 | 91.3 | |
| YOLOV++Backbone=SwinBase, Base detector=YOLOX-SwinB, Time (ms)=15.9, Post-processing=None2024.07 | 90.7 | |
| TransVOD LiteBackbone=SwinBase, Base detector=Deformable DETR, Time (ms)=51.7, Post-processing=None2024.07 | 90.1 | |
| PTSEFormerBackbone=ResNet-101, Base detector=Deformable DETR, Post-processing=None2024.07 | 88.1 | |
| ObjectsBackbone=ResNet-101, Base detector=Deformable DETR, Post-processing=None2024.07 | 87.9 | |
| YOLOV++Backbone=Modified CSP v5, Base detector=YOLOX-SwinT, Post-processing=REPP [46]2024.07 | 86.6 | |
| ClipVIDBackbone=ResNeXt-101, Base detector=Faster R-CNN, Post-processing=None2024.07 | 85.8 | |
| YOLOV++Backbone=SwinTiny, Base detector=YOLOX-SwinT, Time (ms)=8.4, Post-processing=None2024.07 | 85.7 | |
| YOLOV++Backbone=Modified CSP v5, Base detector=YOLOX-L, Post-processing=REPP [46]2024.07 | 85.6 | |
| HVRBackbone=ResNeXt-101, Base detector=Faster R-CNN, Post-processing=Seq-NMS [21]2024.07 | 85.5 | |
| MEGABackbone=ResNeXt-101, Base detector=Faster R-CNN, Post-processing=BLR [9]2024.07 | 85.4 | |
| HVRBackbone=ResNeXt-101, Base detector=Faster R-CNN, Post-processing=None2024.07 | 84.8 | |
| ClipVIDBackbone=ResNet-101, Base detector=Faster R-CNN, Post-processing=None2024.07 | 84.7 | |
| RDNBackbone=RestNeXt-101, Base detector=Faster R-CNN, Post-processing=BLR [9]2024.07 | 84.7 | |
| MAMBABackbone=ResNet-101, Base detector=Faster R-CNN, Post-processing=None2024.07 | 84.6 | |
| MEGABackbone=ResNet-101, Base detector=Faster R-CNN, Post-processing=BLR [9]2024.07 | 84.5 | |
| TROIABackbone=ResNeXt-101, Base detector=Faster R-CNN, Time (ms)=143.2, Post-processing=None2024.07 | 84.3 | |
| YOLOV++Backbone=Modified CSP v5, Base detector=YOLOX-L, Time (ms)=7.6, Post-processing=None2024.07 | 84.2 | |
| SELSABackbone=RestNet-101, Base detector=Faster R-CNN, Post-processing=REPP [46]2024.07 | 84.2 | |
| MEGABackbone=ResNeXt-101, Base detector=Faster R-CNN, Post-processing=None2024.07 | 84.1 | |
| RDNBackbone=RestNet-101, Base detector=Faster R-CNN, Post-processing=BLR [9]2024.07 | 83.8 | |
| HVRBackbone=ResNet-101, Base detector=Faster R-CNN, Post-processing=Seq-NMS [21]2024.07 | 83.8 | |
| TransVOD LiteBackbone=SwinTiny, Base detector=Deformable DETR, Time (ms)=33.8, Post-processing=None2024.07 | 83.7 | |
| YOLOVBackbone=Modified CSP v5, Base detector=YOLOX-L, Time (ms)=6.3, Post-processing=None2024.07 | 83.6 | |
| HVRBackbone=ResNet-101, Base detector=Faster R-CNN, Post-processing=None2024.07 | 83.2 | |
| RDNBackbone=ResNeXt-101, Base detector=Faster R-CNN, Post-processing=None2024.07 | 83.2 | |
| SELSABackbone=ResNeXt-101, Base detector=Faster R-CNN, Time (ms)=76.6, Post-processing=None2024.07 | 83.1 | |
| MEGABackbone=ResNet-101, Base detector=Faster R-CNN, Time (ms)=121.8, Post-processing=None2024.07 | 82.9 | |
| QueryPropBackbone=ResNet-101, Base detector=Sparse R-CNN, Post-processing=None2024.07 | 82.3 | |
| TROIABackbone=ResNet-101, Base detector=Faster R-CNN, Time (ms)=133.5, Post-processing=None2024.07 | 82 | |
| TransVODBackbone=ResNet-101, Base detector=Deformable DETR, Time (ms)=345, Post-processing=None2024.07 | 81.9 | |
| RDNBackbone=ResNet-101, Base detector=Faster R-CNN, Time (ms)=93.1, Post-processing=None2024.07 | 81.8 | |
| SELSABackbone=ResNet-101, Base detector=Faster R-CNN, Time (ms)=80, Post-processing=None2024.07 | 80.3 | |
| YOLOV++Backbone=Modified CSP v5, Base detector=YOLOX-S, Time (ms)=5.3, Post-processing=None2024.07 | 78.7 | |
| FGFABackbone=ResNet-101, Base detector=R-FCN, Post-processing=Seq-NMS [21]2024.07 | 78.4 | |
| YOLOVBackbone=Modified CSP v5, Base detector=YOLOX-S, Time (ms)=4, Post-processing=None2024.07 | 77.3 | |
| FGFABackbone=ResNet-101, Base detector=R-FCN, Time (ms)=62, Post-processing=None2024.07 | 76.3 | |
| EVODBackbone=Modified CSP v5, Base detector=YOLOX-S, Post-processing=None2024.07 | 75.1 |