Using Depth Cues for Enhancing Data Visualization

在这个项目中，我们成功地构建了数学模型来描述透视深度线索对我们对给定可视化图像的视觉注意的影响。结果表明，即使深度线索的相关失真保持在一定的容差范围内，透视图像仍然可以在视觉上保持一致。我们可以进一步提出一种算法，通过在可视化图像中的深度线索的布局中引入过度失真来有意地引导观众的视觉注意。我们的新发现将作为生成高质量可视化图像的基础，这些图像考虑了人类视觉系统固有的特性。

In this project, we successfully constructed mathematical models for describing the influence of perspective depth cues on our visual attention to the given visualization images. The obtained results reveal that perspective images can be still be visually consistent even when the associated distortion of the depth cues stays within in some tolerance. We can further present an algorithm for intentionally directing visual attention from viewers by introducing excessive distortion in the layout of the depth cues in the visualization images. Our new findings will serve as the basis for generating high-quality visualization images that take into account the properties inherent in our human visual system.