EYE LEARN

EYE LEARN is a collaborative project between the School of Health of Fribourg and the School of Engineering and Architecture of Fribourg. It was awarded the 2021 Innovation in Education Prize by the Academic and Pedagogical Development Support Service of HES-SO. This innovative pedagogical process integrates eye-tracking technology and artificial intelligence to enhance the learning experience for Bachelor of Nursing students during patient simulation exercises.

Project Overview

A specially developed web interface, created with Next.js and styled using Tailwind CSS, allows students to review their simulation sessions via video, with the addition of eye-tracking data. Teachers can manage these simulations through the same interface. The system architecture comprises five key components: a web interface, an HTTP server with a REST API, a MySQL database, a Redis instance for task management, and a Python worker for intensive processing tasks.

The REST API, developed in Python, facilitates interaction between the web interface and the system's core data, including user authentication, simulation management, and video playback. All user, simulation, and Machine Learning model outputs are securely stored in a MySQL database. The Redis instance acts as a queue system, handling asynchronous tasks, while the Python worker manages resource-heavy processes like video processing using Machine Learning models trained on the Roboflow platform. These models can be directly imported into the system for analyzing simulation videos.

Technical Infrastructure

The entire system is protected and optimized by a Traefik proxy, which simplifies domain name management and HTTPS certificate handling. Traefik offers additional benefits, including load balancing, dynamic service discovery, and seamless integration with modern container orchestration platforms, ensuring a robust and scalable system architecture.

Impact and Future Development

A total of 114 Bachelor of Nursing students have already tested the EYE LEARN process during five test sessions. The project has resulted in several research papers and has been presented at multiple international conferences. The collaboration between the two institutions will continue, with EYE LEARN set to be further utilized and improved in the future.

Footnotes

I worked on this project as part of my work at the HumanTech Institute.