Modeling Color Difference for Visualization Design
Modeling Color Difference for Visualization Design
复制标题
DOI：
10.1109/tvcg.2017.2744359
复制
发表时间：
2018-01-01
影响因子：
5.2
通讯作者：
Szafir, Danielle Albers
中科院分区：
文献类型：
--
作者：
Szafir, Danielle Albers
颜色经常用于对可视化效果中的值进行编码。要使颜色编码有效，颜色和值之间的映射必须保留数据中的重要差异。然而，可视化中有效颜色选择的大多数指导原则要么基于在最佳观察环境中使用大而均匀的视野测量的颜色感知，要么基于定性直觉。这些限制可能会导致可视化中的数据误解，可视化通常使用细小的细长标记。我们的目标是开发量化指标，帮助人们在可视化中更有效地使用颜色。我们提出了一系列众包研究，测量三种常见标记类型的色差感知：点、条和线。我们的结果表明，人们感知颜色差异的能力在不同的标记类型之间存在显著差异。根据生成的数据构建的概率模型可以为设计者提供客观指导，使他们能够预测观众的感知，以便为有效的编码设计提供信息。
Color is frequently used to encode values in visualizations. For color encodings to be effective, the mapping between colors and values must preserve important differences in the data. However, most guidelines for effective color choice in visualization are based on either color perceptions measured using large, uniform fields in optimal viewing environments or on qualitative intuitions. These limitations may cause data misinterpretation in visualizations, which frequently use small, elongated marks. Our goal is to develop quantitative metrics to help people use color more effectively in visualizations. We present a series of crowdsourced studies measuring color difference perceptions for three common mark types: points, bars, and lines. Our results indicate that peoples abilities to perceive color differences varies significantly across mark types. Probabilistic models constructed from the resulting data can provide objective guidance for designers, allowing them to anticipate viewer perceptions in order to inform effective encoding design.