Researchers have developed a new method called CONF-LA to improve the accuracy of assigning eye-tracking fixations to specific lines of text during reading. This approach integrates knowledge of reading behavior with likelihoods of lines to assign fixations, deferring assignments when uncertainty is high. CONF-LA demonstrates stable performance in post hoc analysis and significantly reduces the gap between online and offline analysis, with a low per-fixation latency of 0.348 ms. It shows particular effectiveness with children's data, improving median accuracies on regressions. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT Improves accuracy and reduces latency in real-time analysis of reading gaze data, potentially aiding educational and accessibility tools.
RANK_REASON Academic paper detailing a new method for analyzing reading gaze data.