When's a Smile a Smile? Or how to Detect a Message by Digitizing the Signal

Abstract

What's in a smile? Why do we respond so powerfully to this visual display? Or, phrased in information processing terms: What characteristics of the signal define the boundaries of the social message? We have used digital image analysis and subject ratings to answer this question for videotaped smiles. A simple measure of the information provided by facial movement—the entropy of the distribution of pixel intensities in the subtracted or difference images—traced the changing facial expression (the signal) through time. Raters categorized the individual videoprints in order to locate message boundaries. We found a remarkable coincidence between changes in signal entropy and message. In each smile, rapid increases in positive messages occurred within entropy crests. We conclude that quantitative measures of information transmission can be used to locate message boundaries and provide insight into how receivers parse the behavioral stream.