Previous efforts such as the USAF Pilot Training Next (PTN) initiative have shown the value of Eye Motion Tracking (EMT) in assisting the trainee to more effectively build required skills when learning where and when to look; how to search, detect, and/or recognize patterns; and how long to gaze. Similar complexity to these DOD operator consoles and cockpits can now be found in myriad applications across many market segments. The purpose of this research paper is to demonstrate the extension of potential benefits of obtaining real-time feedback of trainee eye direction and gaze duration to other use-cases beyond military training, in this case medical imagery interpretation. This real-time indicator can allow the trainer to adapt verbal queueing of the trainee to improve knowledge transfer. It can also provide objective verification evidence that the trainee is indeed looking and gazing at the intended locations on the monitor.
This paper reports on joint Research and Development (R&D) activities exploring Eye Motion Tracking (EMT) for training. EMT is popular in professional online gaming, providing amateur e-sport fans a method of watching precisely where professional players’ eyes are looking as they play complex games with data-intensive user interfaces. The authors are currently evaluating the ability of EMT systems to improve radiology training. Our initial experiment data set includes bone radiographs, digital subtraction angiograms, and axial Computed Tomography (CT) slices. Preliminary results indicate that expert viewers (e.g., board-certified radiologists) were able to use EMT to successfully guide novice readers (e.g., software developers) with no medical training through search and gaze protocol patterns.
We anticipate using human performance data and lessons learned from our radiology training experiments with EMT to identify and investigate other potential training that can benefit from EMT. We will also identify EMT technology enhancements that can benefit the EMT/training experience, including active and passive software-driven queueing, metrics collection, and the use of AI/ML to replicate expert viewer behaviors.
Keywords
Additional Keywords
Eye Motion Tracking, Medical Training