Focus+Context techniques are commonly used in visualization systems to provide both the details and the context. The "shaker paradigm", in which an image is rapidly alternated with a transformed image, was recently developed to find the threshold of perceptual invariance on scaling and rotations. Using this framework, we aim to examine the limits of perceptual invariance on fisheye transformations. We find that there is a "no-cost" region where visual search for a given target is unaffected by an abrupt, noticeable fisheye transformation. There is also a "low-cost" region where transformations of larger magnitude result in performance degradation. We also find that a rectangular grid background neither aids nor impairs performance. Our findings quantify the perceptual cost of nonlinear transformations, confirming that there exist cases where increased context can be experienced with minimal or no cost. These findings contradict the belief that sudden chan...
Keith Lau, Ronald A. Rensink, Tamara Munzner