Eye Tracking Project

This project investigates how readers process texts of varying complexity using eye-tracking data. The goal is to identify objective indicators of reading difficulty and to support the development of accessible written communication.

The project resulted in the FETA (French Eye-TrAcking) corpus, an eye-tracking dataset combining medical, clinical, and general French texts in original and simplified versions.

The FETA Corpus

The corpus contains:

• multiple text types (medical, clinical cases, general texts)
• original and manually simplified versions
• word-level eye-tracking annotations
• participants from different reader populations

Example of Annotated Reading

Below is an example recording showing gaze behaviour during reading. Fixations and scanpaths illustrate how readers allocate attention across the text.

Research Goals

• Detect difficult passages using eye movements
• Model reading effort computationally
• Predict eye-tracking features from text
• Improve medical text accessibility