Eye Tracking: A Promising Means of Tracing, Explaining, and Preventing the Effects of Display Clutter in Real Time.

Eye tracking;Display clutter;Adaptive display;Visual search;Industrial and Operations Engineering;Engineering;Industrial and Operations Engineering

Moacdieh, Nadine MarieLiu, Yili ;

University of Michigan

关键词: Eye tracking;    Display clutter;    Adaptive display;    Visual search;    Industrial and Operations Engineering;    Engineering;    Industrial and Operations Engineering;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/113627/nadmarie_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Display clutter is a widely-acknowledged but ill-defined problem that affects operators in complex, data-rich domains, such as medicine and aviation. Largely regarded a function of data density and display organization, clutter has been shown to degrade performance on a range of tasks, most notably visual search and noticing. Clutter effects may be exacerbated by stress, a major performance-shaping factor in the above domains. The goal of this dissertation was to develop an eye tracking-based approach for tracing and preventing the effects of clutter and stress on attention allocation and information acquisition. The research involved three stages: 1) identify the most diagnostic eye tracking metrics for capturing and explaining the effects of clutter and stress on performance, 2) determine which eye tracking metrics can detect the effects of clutter early on, in real time, and form the basis for models of clutter effects, and 3) evaluate the effectiveness of real-time display adjustments for preventing performance decrements. This research was carried out in several contexts, including emergency department (ED) electronic medical records (EMRs). First, three experiments were conducted in different application domains, including the ED, to establish the relationship between clutter, stress, attention, and performance during visual search and noticing tasks. Clutter resulted in performance decrements on both tasks. The underlying changes in attention allocation were captured by several eye tracking metrics, some of which were able to differentiate between the effects of data density and organization. A fourth experiment calculated the most promising eye tracking metrics in real time and used them as input to logistic regression models of response time. Long response time due to poor organization could be modeled most accurately. Finally, a fifth experiment presented ED physicians with real-time adaptations (highlighting and shortcut panel) to their EMR while they reviewed patient records to perform diagnoses. Both adjustments led to better performance and were viewed favorably by physicians. Overall, this research adds to the knowledge base on clutter and visual attention, supports the further development of eye tracking as a basis for real-time processing, and contributes to improved safety in various domains by supporting timely and accurate information acquisition.