A medical visual feedback system has been developed by a joint research team of NTUEENTUH to assist in inspecting medical images such as X-raysendoscopic films. By presenting the gaze trajectory of experienced physiciansthe critical gaze positions, this AI-enabled system can improve the learning experience of users. The system can also be applied to improve diagnosis efficiencyremind key symptom patterns. The system is inter-operational with various medical platforms, head-mounted displays,eye trackers available on the market.

Unlike previous technologies, our visual feedback system is specially designed for medical applications. By simultaneously considering influential factors about binocular vision, gaze time, symptom size,visual eccentricity, the system accurately generates a visualization of the gazed image areas. It also labels all critical organstissues identified in the visualization process. In addition, our system provides a new way to collect the data required for saliency detection. A saliency map is automatically generated without extra labortime, based on the gaze trajectory of the doctor inspecting the image.

Our visual feedback system targets medical applications. It marks unexamined areas to help reduce the occurrence of negligencethus enhances diagnosis quality. For medical training, our system helps doctors to visualize the gaze trajectorydurationthereby accelerate the learning process of a trainee. In addition, because our system parameters defining the size of recognizable regions are customizable, the system is highly flexibleapplicable to various fields such as driver monitoring, production control, disaster rescue, navigation,flight control.