Jänicke, HeikeChen, MinG. Melancon, T. Munzner, and D. Weiskopf2014-02-212014-02-2120101467-8659https://doi.org/10.1111/j.1467-8659.2009.01667.xSalience detection is a principle mechanism to facilitate visual attention. A good visualization guides the observer s attention to the relevant aspects of the representation. Hence, the distribution of salience over a visualization image is an essential measure of the quality of the visualization. We describe a method for computing such a metric for a visualization image in the context of a given dataset. We show how this technique can be used to analyze a visualization s salience, improve an existing visualization, and choose the best representation from a set of alternatives. The usefulness of this proposed metric is illustrated using examples from information visualization, volume visualization and flow visualization.