Juha Choi
Engineering (electrical)
Hometown: Frisco, Texas, US
Graduation date: Spring 2027
FURI | Fall 2026, Needs Review
Neuro-Symbolic Reasoning for Objective Surgical Skill Assessment
Surgical training traditionally relies on mentorship and expert observation, making skill assessment time-consuming and subject to evaluator variability. This research investigates a neuro-symbolic framework that combines the computational efficiency of neural networks trained on surgical video and instrument motion data with human-controlled, expert-defined symbolic reasoning guided by the Objective Structured Assessment of Technical Skills rubric. The framework connects instrument movement, coordination, and efficiency to interpretable skill assessments. Preliminary experiments suggest that frame shuffling and alternative motion normalization preserve classification accuracy under tested conditions. Future work will refine symbolic reasoning and validate assessments against expert evaluations to support consistent, actionable feedback for residents.
Mentor: Shan Lin