Eye-Tracking Based Automatic Summarization of Lecture Slides

This paper presents a novel approach to automatically summarize lecture slides using eye-tracking data. The tool generates personalized summaries by analyzing learners’ visual attention patterns, with the aim of reducing information overload and enhancing comprehension. While existing text summarization methods often overlook individual preferences and attention patterns, our proposed approach considers all slide elements, including text and visuals. It dynamically selects areas of interest based on learners’ fixation points. The tool is developed as a web application, enabling users to access presentations, receive personalized summaries, and save them for optimized studying. It incorporates features such as administration and settings, slides uploading, and summary viewing. To evaluate the tool, we conducted an experiment involving 62 participants to assess usability and satisfaction with the automated summaries. The evaluation results indicated positive feedback in terms of ease of use and visual appeal. Participants expressed satisfaction with the generated summaries and were given the option to select their preferred summarization percentage. Our analysis revealed that summarization preferences are not related to prior knowledge. Overall, this tool demonstrates the potential of personalized summarization based on eye-tracking data, thereby enhancing learning experiences.