Visual Grounding on RefPath 1.0 (test)
69.95AccuracyPKNet
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PKNetVisual/Text Encoder=RN50/BERT-B2025.02 | 69.95 | 63.49 | |
| D-MDETRVenue=TPAMI'24, Visual/Text Encoder=CLIP-B/CLIP-B2025.02 | 64.92 | 57.69 | |
| SimVGVenue=NeurIPS'24, Visual/Text Encoder=ViT-B/BERT-B2025.02 | 63.94 | 59.42 | |
| LLaVa-MedVenue=NeuIPS'23, Visual/Text Encoder=CLIP-L/LLaMa2025.02 | 62.32 | 57.96 | |
| TransCPVenue=TPAMI'24, Visual/Text Encoder=RN50/BERT-B2025.02 | 61.73 | 56.81 | |
| CLIPVGVenue=TMM'23, Visual/Text Encoder=CLIP-B/CLIP-B2025.02 | 58.89 | 53.97 | |
| TransVGVenue=ICCV'21, Visual/Text Encoder=RN50/BERT-B2025.02 | 58.4 | 52.86 | |
| SeqTRVenue=ECCV'22, Visual/Text Encoder=DN53/BiGRU2025.02 | 55.84 | 51.96 |