Visual quality metrics and human perception

Visual quality metrics have been recently devised to automatically extract interesting visual projections out of a large number of available candidates in the exploration of high-dimensional databases. The metrics permit for instance to search within a large set of scatter plots (e.g., in a scatter plot matrix) and select the views that contain the best separation among clusters. The rationale behind these techniques is that automatic selection of "best" views is not only useful but also necessary when the number of potential projections exceeds the limit of human interpretation. While useful as a concept in general, such metrics received so far limited validation in terms of human perception. In this paper we present a perceptual study investigating the relationship between human interpretation of clusters in 2D scatter plots and the measures automatically extracted out of them. Specifically we compare a series of selected metrics and analyze how they predict human detection of clusters. A thorough discussion of results follows with reflections on their impact and directions for future research.