| Send e-mail to the authors |
Date of initial submission: 5 May 1998
Date of revised/final submission: 8 May 1998
Complex visual data types such as full-motion video, high-resolution images, and interactive animations are becoming more important to end-users as computing power and network bandwidth continue to increase. Information retrieval (IR) systems, previously used exclusively for textual data, have been adapted for visual information retrieval. While these systems are satisfactory solutions for word-based searches, visual information is not easily identified through objective verbal descriptors. For example, different people are likely to use different words to describe the same image. In addition, language is not well suited for describing nuances in visual flow and composition. One solution is to provide users directly with visual information representing a visual document. Like abstracts in text documents, visual abstracts would allow users to obtain a quick overview of the underlying document without having to view it in its entirety.
The purpose of this experiment was to study the effectiveness of two video browsing interface designs (storyboard and slide show) on different video information seeking tasks (gist determination and object recognition). A test system was developed to determine the effects of display type on accurate and rapid completion of specific tasks. The study used a repeated block factorial 2x2 (RBF-22) design; each of the 21 participants interacted with all 4 interface-task combinations. Task accuracy, time to completion, and subjective satisfaction were measured. Although no statistically significant results were found for performance, participants preferred the storyboard interface over the slide show for both tasks.
[ Abstract | Credits | 1. Introduction | 2. Experiment | 3. Results | 4. Discussion | 5. Conclusions | Acknowledgements | References | Appendices ]