Improving Video Navigation for Spatial Task Tutorials by Spatially Segmenting and Situating How-To Videos

How-to videos are widely used for accessing instructional content. Many of the tasks covered in these videos are spatial, requiring movement between locations within a physical space to complete different parts of the activity. Conventional linear video interfaces, which often only allow time-based navigational techniques, like scrubbing, prove inefficient and cumbersome for such tasks. To address this, we investigate an approach for video browsing and navigation optimized for how-to videos involving spatial tasks by chaptering videos based on where tasks occur by using augmented reality to anchor these video segments to their physical locations via virtual signposts. Through two studies, we demonstrate that our approach outperforms standard and chaptered video interfaces in speed and ease. Our work contributes empirical evidence that spatially segmenting and situating tutorials is a promising strategy for improving video navigation.