Date of Award

6-2024

Degree Name

Doctor of Philosophy

Department

Electrical and Computer Engineering

First Advisor

Janos L. Grantner, Ph.D.

Second Advisor

Ikhlas Abdel-Qader, Ph.D.

Third Advisor

Saad A. Shebrain, MD

Fourth Advisor

Robert G. Sawyer, MD

Abstract

For certain surgical procedures, Minimally Invasive Surgery (MIS) has become more advantageous than traditional open surgery. Therefore, mastery of laparoscopic skills is an essential component of surgical training and requires considerable time and effort. The Fundamentals of Laparoscopic Surgery (FLS) program has been developed as a tool to improve and assess fundamental surgical skills. In fact, using a low-cost Box-Trainer or a simulator, laparoscopic surgeons are required to train using a set of structured tasks that can be objectively used to assess their laparoscopic skills, which must be mastered before carrying out real-life laparoscopic procedures. These tasks include peg transfer, clip and divide, pattern-cutting, suturing with intracorporeal or extracorporeal knots, mesh placement, and fixation. To guarantee that the use of laparoscopic surgical training methods results in the mastery of essential surgical skills, the current approach requires observation by expert medical personnel which is time-consuming and not cost-effective. To address these challenges and minimize human supervision, an automated skill assessment system is essential.

In this dissertation, to assess surgeons’ skill level in peg transfer and pattern-cutting tasks by using our Intelligent Box-Trainer System (IBTS) platform, several automatic skill assessment systems have been proposed. These systems can detect laparoscopic instruments and track the surgeon’s hand motion. For this purpose, a first-of-its-kind, custom laparoscopic box-trainer dataset was built from experimental peg transfer and pattern-cutting video recordings which were carried out by the help of 9 doctors and OB/GYN residents, of the Homer Stryker M.D. School of Medicine, WMU, in the Intelligent/Fuzzy Controllers Laboratory, at WMU. The proposed methods are based on the Tensorflow Object Detection API in which the feature extractor was trained on the new custom dataset. For the pattern-cutting task, the tooltip tracking has been implemented using two different automated skill assessment methodologies. In the first one, a basic trainees’ skill assessment system is designed, based on several If-Then statements, depending on tooltip distance measurements from the center of the test platform, and a set of metrics, corresponding with the evaluation criteria. In the second method, a Multi-Input-Single-Output (MISO) Fuzzy Logic Supervisor system has been applied to the measured tooltip distance to assess the surgeons’ skill levels in percentages. As for the peg transfer task, two automatic skill assessment methods have been proposed. In the first method, an automated sequential assessment algorithm has been developed, based on several sequential If-Then conditional statements, and by monitoring the surgeon’s performance by using top, side, and front cameras and determining whether the surgeon executes all the peg transfer task procedures in a hierarchical fashion. In the second method, to assess the surgeons’ left and right-hand movement in 3D space, an automated evaluation system based on a multi-thread video processing application method has been proposed. Its operation is based on laparoscopic instrument detection with respect to the target platform, and a cascaded fuzzy logic assessment system. After implementing the algorithm on the Intelligent Box Trainer System (IBTS), the validity of the proposed methodologies has been examined on both recorded videos and under real-time IBTS operation. However, due to inherent delays observed under real-time conditions, , we plan to improve the computing power of the IBTS (e.g., by adding an NVIDIA GeForce RTX 3080) to reduce this delay, in our future work.

For supporting research in the Laparoscopic Surgery Skill Assessment area, our custom dataset has been made available, free of charge, at https://drive.google.com/drive/folders/1F97CvN3GnLj-rqg1tk2rHu8x0J740DpC, and the code will be shared in GitHub at https://github.com/Atena-Rashidi).

Access Setting

Dissertation-Open Access

Download

Included in

Biomedical Commons

Share

COinS