Development of a machine learning model for automatic assessment of performance in virtual reality medical simulators
Virtual reality simulators for medical training allow the student to learn and train medical procedures safely, before performing on a real patient. By collecting data from the trainees while they perform in the simulator, it is possible for the system to automatically give feedback about the student’s performance in the procedure, in real time, making the training more independent, while also yielding information to the instructor that can be insightful regarding their student’s learning. As studies using machine learning techniques to automatically assess performance in virtual reality simulators are scarce in the literature, this work proposes a model by comparing the results of different classifiers by using data collected in a dental simulator. The results so far have shown that the Naive Bayes classifier has achieved a lower Accuracy in most cases when compared to the Support Vector Machine, Multi-layer Perceptron and Random Forest classifiers.
S. M. B. I. Botden, I. H. J. T. de Hingh, and J. J. Jakimowicz. 2009. Suturing Training in Augmented Reality: Gaining Proficiency in Suturing Skills Faster. Surgical Endoscopy 23, 9 (Sept. 2009), 2131–2137.
James Brewin, Tim Nedas, Ben Challacombe, Oussama Elhage, Jonas Keisu, and Prokar Dasgupta. 2010. Face, Content and Construct Validation of the First Virtual Reality Laparoscopic Nephrectomy Simulator. BJU International 106, 6 (Sept. 2010), 850–854.
Cleber Gimenez Correa, Romero Tori, and Fatima L. S. Nunes. 2013. Haptic Simulation for Virtual Training in Application of Dental Anesthesia. In 2013 XV Symposium on Virtual and Augmented Reality. IEEE, Cuiabá - Mato Grosso, Brazil, 63–72.
Romero Tori, Gustavo Ziyu Wang, Lucas Henna Sallaberry, Allan Amaral Tori, Elen Collaço de Oliveira, and Maria Aparecida A. M. Machado. 2018. VIDA ODONTO: Ambiente de Realidade Virtual Para Treinamento Odontológico. Revista Brasileira de Informática na Educação 26, 02 (May 2018), 80–101.