Experimental study on perception of color encoding based on information interface statistical task

GUO Qi,WU Jin-chun,XUE Cheng-qi

DOI: https://doi.org/10.11996/jg.j.2095-302x.2023020408

2023-01-01

Abstract:In order to effectively enhance users’ decision-making efficiency for color encoding in visual statistics tasks,it is necessary to first investigate the perceptual problem of the human visual system for color encoding in visualization interfaces. Therefore, the concept of significant difference was introduced into the color encoding research, leading to the creation of a set of new color spaces for visual statistics tasks. The effect of color difference threshold on the perceptual performance of visual statistics tasks was then examined through typical visual statistics task experiments, so as to summarize the perceptual accuracy and laws of color coding for visual statistics tasks. The results revealed that: firstly, in the correlation recognition task and the mean comparison task, the perceptual accuracy of the human visual system was more accurate than that predicted by previous color models, indicating that the difference threshold（JND） selection for color encoding could be smaller. Secondly, when different color axes were selected for color encoding, the recognition performance of the correlation recognition task was higher with lightness encoding. Finally, under different basic correlation coefficients and number of point groups, the accuracy of correlation recognition and mean comparison followed the Weber linear function, which reflected the observers’ability to process information entropy when performing visual statistical tasks. The above findings provide design guidance for information visualization designers, and offer a quantitative basis for color coding of visual interface information from the perspective of psychophysics.