Abstract
Video comparison is a common task that people engage in when interacting with videos. A learner improves their performance in video-based learning by comparing themselves to a good model; a producer composes a compelling video by comparing clips and selecting the best one during media production; many domain areas and applications benefit from the comparison task. However, the comparison of videos is not easy because of the rich content and unique temporal characteristics of a video. This issue raises the need for special considerations when designing the user interface to facilitate the comparison. One important aspect of the considerations is complexity reduction, where a system turns videos and their relationships into more abstract or simpler forms. For instance, a system may highlight parts that need attention or extract some key events as an image sequence to ease the comparison. This thesis synthesizes existing comparison knowledge and applications into a design space, then applies the design space to video-based learning and video production applications. I used an iterative design process and conducted user studies with five different applications: storyboard, debugger, physical training, video analysis, and live streaming. These applications demonstrate the usefulness of complexity reduction approaches (element highlighting, transforming, and partitioning) for each video component (including spatial component, temporal component, and relationships between videos) in facilitating comparison tasks. The findings show that complexity reduction is an essential design element for facilitating visual comparison of videos, and considering how to apply complexity reduction to a video component can generate a novel user interface that benefits usability and facilitates video comparison. This thesis offers a promising basis for a conceptual framework when designing video comparison applications.