Estimation of the capacity of human perception using computational aesthetics

Abstract

The paper is part of a broader study to utilize some hypotheses, methodologies and tools from Computational Aesthetics, the foundations of which were created in C.R. Shannon's »The Mathematical Theory of Information». Using the basic ideas of computational aesthetics, a picture can be seen as a message. Cybernetic aesthetics introduces information flow to describe perceptible information. Further cybernetic aesthetic proposes characteristics for conspicuity and intelligibility of artwork. In the sample survey, using these variables, we try to measure the quantity of perceptible information in the image. In the literature, human perceptual capacity values are reported close to 15 bits per second. Our aim will be to verify this information. We will do our study with 200 responses of 35 respondents, where we will measure their ability to perceive different elements in M.C. Escher's "wallpaper" paintings. Part of the study will be a statistical evaluation of the measurement.