Answers are grounded in this paper and cite their sources.
Manikin-based medical training is challenging and costly, traditionally requiring highly trained professionals to provide detailed user interaction assessments and feedback. In addition, colonoscopy training has a steep learning curve due to the complexity of colonoscope control, where an endoscope can loop inside the body, requiring specialized movements to resolve. The ability to recognize when a loop is forming is key to effective training. This study presents a hybrid colonoscopy training system designed to detect loop formation during manikin-based training; thereby providing automated assessment and feedback to the user. This work addresses current limitations in existing training systems by improving the ability to detect loop formation across different patient anatomies. The system integrates electromagnetic position tracking and force sensing with a machine learning-based loop detection algorithm to assess user interactions. Real-time feedback is delivered through a unity-designed interface. A human study demonstrated that the system can accurately distinguish differences in insertion force and procedure time between expert and novice users. In addition, three machine learning algorithms were evaluated, with the transformer model achieving the most robust performance, reaching an average real-time loop detection accuracy of 93.3%. This approach enhances colonoscopy training by providing automated, objective assessment and real-time feedback, reducing reliance on expert supervision and improving training accessibility while promoting safer procedural techniques.
No takes yet. Share an insight, caveat, or question.
Wu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: