Research Talk: From Perceptual Model of Visualization to Visual-Centric Computation
In the era of big data, computation of big data is extremely expensive and even impractical. When visualizing the results of computation of big data, visual limitations from both the screen resolution and human perception exacerbate the difficulty of data analysis and further restrict the understanding of the information beyond the data. One of the possible approaches, which is called visual-centric computation, is to use the visual limitations to guide the computation and the rendering of the visualization.
In this talk, Yang will present her research towards visual-centric computation. First, she will introduce our existing work on quantitatively modeling the perception of correlation in bivariate visualization, using a classic perceptual model called Weber’s law. Then she will present her on-going work on analyzing visual features within the visualizations, to understand why the Weber’s law holds for the perception of correlation. Finally, Yang will discuss how visual features and perceptual models can be used to guide the visual-centric computation.