Yes/No Medical Visual Question Answering on VinDr
72.8F1 ScoreLobA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LobAModel Category=Adversarial Techniques2025.04 | 72.8 | 75.8 | 71.1 | |
| LobA w/o self-promptModel Category=Adversarial Techniques, Self-prompt=false2025.04 | 72.4 | 74.5 | 70.6 | |
| PH + LLaVA-MedModel Category=Adversarial Techniques2025.04 | 71.6 | 72.5 | 69.8 | |
| LISAModel Category=Open-source Models2025.04 | 70.7 | 71.2 | 69.4 | |
| LLaVA-MedModel Category=Open-source Models2025.04 | 70.3 | 70.8 | 69.9 | |
| CRG + LLaVA-MedModel Category=Adversarial Techniques2025.04 | 68.9 | 66.7 | 70.2 | |
| VCD + LLaVA-MedModel Category=Adversarial Techniques2025.04 | 65.2 | 69.9 | 59.1 | |
| BioMedGPTModel Category=Open-source Models2025.04 | 62.4 | 60.6 | 63.1 | |
| CheXagentModel Category=Open-source Models2025.04 | 61 | 62.8 | 60.3 | |
| GPT-4-OModel Category=Proprietary Models2025.04 | 59.9 | 66 | 48.7 | |
| MedFlamingoModel Category=Open-source Models2025.04 | 58.4 | 60 | 54.7 | |
| Gemini-1.5-Flash-8BModel Category=Proprietary Models2025.04 | 57.1 | 60.4 | 54.3 | |
| GPT-4-VisionModel Category=Proprietary Models2025.04 | 56.8 | 63.7 | 42.4 |