A Virtual Coaching Approach For the Computer-Assisted Surgical Trainer

This study proposes a novel virtual coaching approach integrated into a computer-assisted surgical trainer to enhance simulated practice for minimally invasive surgical procedures. The proposed system utilizes a neural network to automate performance evaluation and feedback, thus addressing the limitations of subjective assessments. Two datasets - one derived from prior research and another newly generated - demonstrated the neural network's capacity to provide accurate grading and targeted recommendations for improvement. While the results highlight the potential of AI-driven systems to standardize evaluations, challenges such as overfitting in smaller datasets underline the need for more diverse data and advanced models in future research.