Group Activity Recognition on VBD (test)
93.9MCAGroup-DINOmics
Evaluation results
| Method | Links | |
|---|---|---|
| Group-DINOmicsExtractor=DINOv3, Input Type=person bounding boxes + whole image, Fine-tuning=true2026.04 | 93.9 | |
| Flaming-NetExtractor=Inception-v3, Input Type=whole image, Fine-tuning=true2026.04 | 93.3 | |
| SOGARExtractor=ViT-Base, Input Type=whole image, Fine-tuning=true2026.04 | 93.1 | |
| MP-GCNExtractor=YOLOV8x, Input Type=person bounding boxes + whole image, Fine-tuning=true2026.04 | 92.8 | |
| KRGFormerExtractor=Inception-v3, Input Type=person bounding boxes + whole image, Fine-tuning=true2026.04 | 92.4 | |
| DFWSGARExtractor=ResNet-18, Input Type=whole image, Fine-tuning=true2026.04 | 90.5 | |
| SAMExtractor=ResNet-18, Input Type=person bounding boxes + whole image, Fine-tuning=true2026.04 | 86.3 | |
| FAGARExtractor=YOLOV8x, Input Type=person bounding boxes + whole image, Fine-tuning=true2026.04 | 85.2 | |
| LiGARExtractor=ResNet-18, Input Type=whole image, Fine-tuning=true2026.04 | 74.8 |