University of Illinois at Urbana-Champaign, USA.
Hum Factors. 2011 Apr;53(2):142-53. doi: 10.1177/0018720811404026.
The present study developed and validated a stochastic model of overt attention within a visual workspace.
Technical specifications and recommended practices for the design of visual warning systems emphasize the role of alert salience. Task demands and display context can modulate alert noticeability, however, meaning that salience alone does not guarantee attention capture.
A stochastic model integrated elements from existing models of visual attention to predict attentional behavior in dynamic environments.Validation studies tested the predictions of the new model against scanning data from a high-fidelity simulator study and behavioral data from an alert detection experiment.
The model accurately predicted the steady-state distribution of attention within a simulated cockpit as well as the effects of color similarity, eccentricity, and dynamic visual noise on miss rates and response times in the alert detection task.
The model successfully predicts attentional behavior in complex visual workspaces with the use of parameter values selected by either the modeler or a subject matter expert.
The model provides a tool to test the effectiveness of visual alerts in various display configurations and with varying task demands.
本研究开发并验证了一种视觉工作空间中显性注意的随机模型。
视觉警示系统的设计技术规范和推荐实践强调了警示显著性的作用。然而，任务需求和显示上下文可以调节警示的可见性，这意味着仅仅显著性并不能保证注意力的捕捉。
一个随机模型集成了现有视觉注意模型的元素，以预测动态环境中的注意力行为。验证研究根据高保真模拟器研究中的扫描数据和警报检测实验中的行为数据，测试了新模型的预测。
该模型准确地预测了模拟驾驶舱内注意力的稳态分布，以及颜色相似性、偏心度和动态视觉噪声对警报检测任务中错误率和响应时间的影响。
该模型使用模型构建者或主题专家选择的参数值，成功地预测了复杂视觉工作空间中的注意力行为。
该模型为在各种显示配置和不同任务需求下测试视觉警报的有效性提供了一种工具。
