Surgical Video Question Answering on EndoVis18-VQA Out-of-template

28.15BLEU-4

RL framework over Digital Twin Representations

Updated 1mo ago

Evaluation Results

Method
RL framework over Digital Twin Representations 2026.06	28.15	55.9	68.45	53.4
SurgViVQA 2025.11	24.84	50.86	50.13	9.23
SurgViVQA 2026.06	24.84	50.86	50.13	9.23
PitVQA 2025.11	17.57	46.87	45.49	9.73
PitVQA 2026.06	17.57	46.87	45.49	9.73
Qwen3-VL 2026.06	15.8	55.2	66.9	50.3
Surgical-LVLM 2026.06	15	54.8	66.2	49.8
Qwen3-VL 2026.06	13.2	53.5	64.8	48.1
Qwen2.5 2025.11	11.92	52.77	63.72	46.9
Qwen2.5-VL 2026.06	11.92	52.77	63.72	46.9
MedGemma 2025.11	7.74	40.74	61.59	37.86
MedGemma 2026.06	7.74	40.74	61.59	37.86
InternVL3 + LoRA 2025.11	6.27	28.04	37.08	56.42
SurgicalGPT 2025.11	2.52	43.97	44.99	11.85
SurgicalGPT 2026.06	2.52	43.97	44.99	11.85
SurgViVQA 2025.11	0.6	12.33	8.1	0.99
InternVL3 2025.11	0.57	38.73	38.73	21.99
VideoLLaMA3 2025.11	0.57	38.73	39.51	35.89
InternVL3 2026.06	0.57	38.73	38.73	21.99
VideoLLaMA3 2026.06	0.57	38.73	39.51	35.89