Referring Expression Comprehension on gRefCOCO (val)
62.1Precision (F1=1, IoU>=0.5)SimVG-DB
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SimVG-DBVisual Encoder=ViT-B/32, Textual Encoder=N/A2024.09 | 62.1 | 54.7 | |
| SimVG-TBVisual Encoder=ViT-B/32, Textual Encoder=N/A2024.09 | 61.3 | 56.1 | |
| UNINEXTVisual Encoder=RN50, Textual Encoder=BERT2024.09 | 58.2 | 50.6 | |
| MDETRVisual Encoder=RN101, Textual Encoder=RoBERTa2024.09 | 42.7 | 36.3 | |
| MM-GDINOBackbone=Swin-T, Zero-shot=true2025.01 | 41 | 91.3 | |
| LLMDetBackbone=Swin-T, Zero-shot=true2025.01 | 40.7 | 91.2 | |
| MM-G-T(c3)Backbone=Swin-T, Setting=zero-shot2024.01 | 40.6 | 83.1 | |
| G-DINO-T(c)Backbone=Swin-T, Setting=zero-shot2024.01 | 40.5 | 83.8 | |
| Grounding-DINOBackbone=Swin-T, Zero-shot=true2025.01 | 40.5 | 83.8 | |
| VLTVisual Encoder=DN53, Textual Encoder=GRU2024.09 | 36.6 | 35.2 | |
| MCNVisual Encoder=DN53, Textual Encoder=GRU2024.09 | 28 | 30.6 |