Department of Software Engineering
Permanent URI for this collection
Browse
Browsing Department of Software Engineering by Author "ÇAĞILTAY, Nergiz"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
Item IMPLEMENTATION OF MACHINE LEARNING METHODS TO UNDERSTAND SURGICAL RESIDENTS' SKILL LEVELS THROUGH THEIR HAND MOVEMENTS GENERATED BY COMPUTER-BASED SIMULATION TRAINING ENVIRONMENTS(2023-07) TONBUL, Gökçen; ÇAĞILTAY, Nergiz; TOPALLI, DAMLAMedical disciplines have been experiencing big challenges in its existing complex nature, parallel with the development of the new technologies. Classical approaches evolve into modern solutions in the adaptation process even some are becoming completely obsolete. The natural complications of an ordinary open surgery directed this evolution towards the term minimally invasive operations. Minimally invasive surgery (MIS), as a general term, uses or creates cavity in the body to reach the desired body part by using necessary tools. The aim is to give less pain to the patient by keeping less incision and tissue damage. However, there are still several problems for the education programs of related surgical procedures. For instance, defining and objectively measuring the surgical skill levels is a challenging process. In this regard, first a systematic review study is conducted to better understand the surgical skill level classification approaches. Afterwards, it is aimed to classify intermediate and novice surgical skills with higher accuracy compared to the previous classification efforts using any possible hand movement-oriented data gathered through virtual reality environments in an experimental study. The results show that it is possible to improve the classification more using different data engineering techniques based on a reproducible adapted framework. It is believed that, in the future, it is possible to adapt this research study effort to any virtual environment with a proper set of tools, the applicable software engineering efforts on top of data science discernment, as well as possible innovative machine learning approximations.