Understanding reader engagement is critical for webtoon creators, yet existing methods rely heavily on indirect signals such as click-through rates, overall dwell time, and reader comments. These metrics provide only a partial and indirect view of rea...
Understanding reader engagement is critical for webtoon creators, yet existing methods rely heavily on indirect signals such as click-through rates, overall dwell time, and reader comments. These metrics provide only a partial and indirect view of readers’ actual attention and engagement during reading. In response to this challenge, we present EYETOON, a content-aware visualization system that helps webtoon creators analyze reader engagement through webcam-based eye-tracking data and LLM-assisted interpretation. By automatically detecting notable reading behaviors and linking them to specific visual features and inferred interpretations, EYETOON offers clear, interpretable and actionable insights into reader engagement disruptions. In a user study with domain experts, participants reported that the system effectively supports them in interpreting engagement patterns and offering insights not accessible through traditional feedback. EYETOON helps creators make data-informed revisions to narrative flow, layout, and character development.