Abstract
The analysis of eye tracking data facilitates the discovery of what people are interested in. Such information can help to improve the design of advertisements or suggest proper products. However, some issues are still left. We focus on the following two tasks: the visualization of long transition patterns among areas of interests (AOIs) and the comparison of scan-paths. In this paper, we propose a new visualization technique to facilitate these tasks. First, we define hierarchical AOIs and convert the scan-paths to strings. We then split these strings into N-grams for the extraction of behavior pattern. Finally, we visualize the results and trajectories of selected patterns. We apply our technique to example cases of scan-paths on several static stimuli. The results show our technique is effective to find patterns of transitions and differences of behaviors.
