Role Purpose

We are looking for one experienced candidate to join our team at the Division of Imaging Science and Technology and Dundee Institute for Healthcare Simulator at the School of Medicine led by Dr Benjie Tang, and Professor Mustafa Suphi Erden at the School of Engineering & Physical Sciences, Heriot-Watt University, to work on a project titled "Autonomous and Intelligent Laparoscopy (AILap) Trainer with Real-Time

Feedback" with the aim of developing a next-generation laparoscopy training platform by bringing together physical realism, digital computation, and developing automated real-time feedback along with the assessment.

This vacancy arises from a grant awarded by the Engineering & Physical Sciences Research Council (EPSRC) to a collaborative project between the University of Dundee and Heriot-Watt University. We will develop a fully autonomous laparoscopy self-training system that provides insightful, intuitive, and timely feedback for improvement. We will develop machine vision and machine learning techniques and integrate those with physical box trainers. Our techniques will also be compatible with virtual reality trainers to equip them with automated immediate feedback. The AILap system will undergo testing on a commonly performed laparoscopic procedure, specifically focussing on proficiency gain and the learning curve associated with laparoscopic suturing skills.

PDRA

The successful candidate will be involved in the following tasks: 1) identifying the essential steps and techniques of laparoscopic suturing through a dedicated study, aiming to define a refined set of steps with a specific focus on when, why, and how real-time feedback would be necessary by achieving a consensus of our Expert Panel (EP); 2) systematically recording videos of approximately 20 trainee and 10 expert surgeon performances in a full laparoscopic suturing exercise using standard laparoscopy training boxes in the surgical skills training centre in Dundee; 3) extracting instrument trajectories and labelling their segments according to the skill levels assigned to the procedures in the videos using an established algorithm. These trajectories will be utilised with machine learning algorithms to develop our AI for detecting movements, identifying skill levels, and deciding on which expert video to demonstrate as feedback when an unskilled-type movement is detected; 4) assessing the performances of 10 expert and 20 novice surgeons, evaluated in agreement by two experts; 5) collecting systematic feedback through qualitative interviews and quantified questionnaires after each user study; 6) utilising questionnaires and interviews for both early and late versions to validate the effectiveness and usefulness of the system for training; 7) collaborating with the team at Heriot-Watt University on various tasks within the same project.