Sure About That Line? Approaching Confidence-Based, Real-Time Line Assignment in Reading Gaze Data

2026-04-27

AI Summary

This work addresses the unreliability of real-time line assignment for gaze points in multi-line reading using remote eye tracking, which stems from noise and layout ambiguity. The authors propose CONF-LA, a novel method that integrates reading behavior priors with Gaussian line likelihoods of gaze points through Bayesian posterior inference to compute line scores. Crucially, CONF-LA introduces, for the first time, a confidence-driven deferred assignment mechanism that postpones decisions under high uncertainty, thereby accommodating natural reading behaviors such as regressions. Experimental results demonstrate that CONF-LA achieves a median accuracy of approximately 95% on public datasets, with an average per-fixation latency of only 0.348 milliseconds. The gap between online and offline performance is reduced to 1–2%, and the method exhibits strong robustness on children’s data.

Abstract

Remote and webcam-based eye tracking in multi-line reading suffers from various noise factors and layout ambiguity, precisely where real-time reading support needs reliable, per-fixation line assignment. Prior work largely addresses this challenge post hoc or by restricting behavior (e.g., disallowing re-reading), undermining interactive use. We propose CONF-LA (Confidence-score-based Online Fixation-to-Line Assignment), a principled, low-latency approach that integrates knowledge about reading behavior and Gaussian line likelihoods over fixations to compute a posterior-line-score and defers assignments when uncertainty is high. Evaluated on existing open-source data, CONF-LA demonstrates stable performance in post hoc analysis and closes the online-offline gap (1-2 %) with a mean per-fixation latency of 0.348 ms. Our approach exhibits particular invariance toward regressions, yielding significant improvement in ad hoc median accuracies on children data (approx. 95 %) over all tested algorithms. We encourage further research in this direction and discuss possibilities for future development.

Problem

• eye tracking
• line assignment
• reading gaze data
• real-time
• noise

Innovation

• confidence-based assignment
• real-time line assignment
• eye tracking
• reading gaze data
• Gaussian line likelihood