Gaze information channel

Abstract

This paper presents the gaze information channel. Using the fact that eye tracking fixation sequences through areas of interest (AOI) have been previously modeled as a Markov chain, we go further by using the fact that a Markov chain is a special case of information, or communication, channel. Thus, the Markov chain quantities of entropy of the equilibrium distribution and conditional entropy are enriched for interpretation of the gaze sequences with the additional quantities of the information channel, such as joint entropy, mutual information and conditional entropy of each area of interest. We illustrate the gaze channel with several examples using Van Gogh paintings. Our preliminary results show that the channel quantities can be given a coherent interpretation to both classify the observers and the artworks.