Modeling surgeon-specific instrument usage profiles to improve instrument anticipation

Anticipating surgical instrument usage is essential for robotic assistance systems to facilitate smooth intraoperative workflows. This study investigates whether modeling surgeon-specific profiles improves the prediction of laparoscopic instrument sequences. We evaluate a profile-aware GRU against a standard GRU and a hidden Markov model. The results demonstrate that incorporating surgeon-specific profiles enhances predictive performance, highlighting the importance of personalization for effective robotic assistance in surgery.