GazeGraphVis: Visual analytics of gaze behaviors at multiple graph levels for path tracing tasks
GazeGraphVis: Visual analytics of gaze behaviors at multiple graph levels for path tracing tasks
复制标题
DOI：
10.1016/j.aei.2023.102111
复制
发表时间：
2023-08
期刊：
影响因子：
--
通讯作者：
Zhuo Yang;Yaqi Xie;Ming Li;George Q. Huang
中科院分区：
文献类型：
--
作者：
Zhuo Yang;Yaqi Xie;Ming Li;George Q. Huang
图形可视化有助于有效地理解图形数据中相互关联的属性。然而，指数级增长的互连对图形数据的高效可视化认知提出了巨大挑战。表达性图形可视化的生成需要研究图形可视化的认知过程，这可以通过对凝视行为的分析来揭示。在本文中，我们提出了视觉分析系统GazeGraphVis来分析路径追踪任务的凝视行为。具体来说，GazeGraphVis通过概览+细节技术在图、节点和边的多个层次上可视化注视行为，以提供对人类完成路径跟踪任务时的认知过程的全面分析。路径跟踪任务的凝视行为的见解揭示使用集成的多视图界面。可视化和眼动分析领域的专家对GazeGraphVis能够获得整体搜索趋势并深入分析影响注视行为的因素给予了高度评价。
Graph visualization contributes to an efficient understanding of interconnected properties in graph data. However, the exponential growth of interconnections poses great challenges to the efficient visual cognition of graph data. Generation of expressive graph visualization requires investigations of the cognitive process of exploring graph visualization, which can be revealed through the analysis of gaze behaviors. In this paper, we propose GazeGraphVis, a visual analytics system, to analyze gaze behaviors for path tracing tasks. Specifically, GazeGraphVis visualizes gaze behaviors at multiple levels of graph, node and edge with overview + detail techniques to provide a comprehensive analysis of human cognitive processes when finishing path tracing tasks. The insights of gaze behaviors for path tracing tasks are revealed using an integrated multiple-view interface. Domain experts in visualization and eye tracking analysis gave high praise to GazeGraphVis for its capability of obtaining the overall search tendencies and deeply analyzing the factors that affect gaze behaviors.