Strabismus Subtype Classification on Strabismus Subtype Classification Dataset
1.44ConsistencyQwen + Reflexion
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Qwen + ReflexionMethod Category=Prompt-based, Base LVLM Model=Qwen, Prompting/Reasoning Framework=Reflexion2026.06 | 1.44 | 1.35 | 1.37 | 43.9 | 40.2 | 42.3 | 15 | 13.6 | |
| GPT + CoVeMethod Category=Prompt-based, Base LVLM Model=GPT, Prompting/Reasoning Framework=CoVe2026.06 | 1.84 | 1.87 | 1.8 | 60.7 | 45.8 | 43.1 | 26.2 | 30.7 | |
| Qwen + CoVeMethod Category=Prompt-based, Base LVLM Model=Qwen, Prompting/Reasoning Framework=CoVe2026.06 | 1.85 | 1.98 | 1.79 | 63.6 | 23.4 | 32.4 | 17.8 | 11.1 | |
| GPT + ReflexionMethod Category=Prompt-based, Base LVLM Model=GPT, Prompting/Reasoning Framework=Reflexion2026.06 | 2.03 | 2 | 1.98 | 70.1 | 41.1 | 74.9 | 38.3 | 41.5 | |
| Gemini-3-flash-preview + ReflexionMethod Category=Prompt-based, Base LVLM Model=Gemini-3-flash-preview, Prompting/Reasoning Framework=Reflexion2026.06 | 2.12 | 1.99 | 2.06 | 65.4 | 16.8 | 2.3 | 15 | 4 | |
| HealthGPTMethod Category=Fine-tuning-based2026.06 | 2.15 | 1.99 | 2.05 | 51.4 | 70.1 | 42.8 | 35.5 | 38.8 | |
| HuatuoGPT-VisionMethod Category=Fine-tuning-based2026.06 | 2.16 | 2.08 | 2.05 | 56.1 | 64.5 | 36.5 | 31.8 | 33.9 | |
| Gemini-3-flash-preview + CoVeMethod Category=Prompt-based, Base LVLM Model=Gemini-3-flash-preview, Prompting/Reasoning Framework=CoVe2026.06 | 2.43 | 2.33 | 2.45 | 71 | 18.7 | 79.1 | 16.8 | 9.6 | |
| Qwen + MAGISMethod Category=Ours, Base LVLM Model=Qwen, Prompting/Reasoning Framework=MAGIS2026.06 | 3.42 | 3.35 | 4.2 | 94.4 | 88.8 | 83.3 | 83.2 | 79.3 | |
| GPT + MAGISMethod Category=Ours, Base LVLM Model=GPT, Prompting/Reasoning Framework=MAGIS2026.06 | 3.64 | 3.95 | 3.9 | 93.5 | 93.5 | 84.5 | 87.9 | 85.2 | |
| Gemini-3-flash-preview + MAGISMethod Category=Ours, Base LVLM Model=Gemini-3-flash-preview, Prompting/Reasoning Framework=MAGIS2026.06 | 4.52 | 4.14 | 4.41 | 95.3 | 97.2 | 91.1 | 92.5 | 91.3 |