Search

This works presents a design and development of Virtual Rotator Cuff Arthroscopic Skill Trainer (ViRCAST) and its preliminary subject study analysis using machine learning approach. Arthroscopy is a minimally invasive surgical intervention regarded as a part of orthopedic sub-specialty. The procedures are performed via small incisions in the patient's skin to examine, diagnose and repair the injuries inside a joint [1]. Surgeons insert tiny instruments and small lens and lighting (called arthroscope) into the joint. They perform surgical intervention seeing the anatomy on a 2D monitor screen streamed from arthroscope. Due to non-natural hand-eye coordination, narrow field-of-view and limited instrument control, training for arthroscopy is challenging and difficult to master. In this work, we developed a primarily ViRCAST platform for training the shoulder arthroscopy procedures. We performed initial validation study using 10 surgery resident subjects (Post-Graduate Year (PGY) 1-5) and performed statistical analysis to extract significant data features. This is followed with machine learning algorithms to cluster and classify the subject's expert level with training data. Our results show that we could successfully distinguish the expertise level.

Data collected and generated through and posterior to projects, such as data residing in project management software and post-project review documents, can be a major source of actionable insights and competitive advantage. This paper presents a rigorous methodological analysis of the applied research published in academic literature, on the application of data mining (DM) for project management (PM). The objective of the paper is to provide a comprehensive analysis and discussion of where and how data mining is applied for project management data and to provide practical insights for future research in the field.