TOWARDS ATTENTION-AWARE CONCEPT MAP BASED REVIEW IN VIDEO LEARNING

Abstract

Video review is commonly performed by learners to clarify or make up for the content they have missed. However, it is not well-supported in the current video learning environment. This thesis covers the design exploration, empirical evaluation, and technical development of CMReview, a technique that for facilitating effective video review. Our observational study identifies three design goals: overview support, navigation support, and attention-aware. Two video review techniques, CMReview and TSReview, were designed and compared. CMReview was proved more effective for understanding long videos and received higher user preference. We then focused on the technical development of CMReview, which involves creating concept maps from video content and developing the attention-aware component. The ScaffoMapping tool was introduced to assist users in creating high-quality concept maps, while a novel moment-to-moment attention detection technique was proposed for annotating learner attention states. This thesis fills a gap in improving the video review experience.