Surgical-LVLM adapts large vision-language models to answer complex visual questions during robotic surgical procedures with unprecedented accuracy and grounding capability.

• Superior medical reasoning through specialized surgical knowledge integration
• Enhanced visual grounding that precisely identifies relevant anatomical structures
• Improved contextual understanding enabling more natural surgeon-AI interactions
• Demonstrated effectiveness on standard EndoVis datasets

This research represents a significant advancement for medical robotics by enabling AI systems to better understand surgical environments, potentially reducing surgical errors and improving patient outcomes through real-time visual guidance and automated surgical mentorship.

Original Paper: Surgical-LVLM: Learning to Adapt Large Vision-Language Model for Grounded Visual Question Answering in Robotic Surgery