A comparison of attention estimation techniques in a public display scenario

Abstract

Human interaction with a public display presupposes a person’s attention. An Interactive display, hence, aims at attracting attention by e.g. emitting a strong signal that makes the inattentive visitor turn towards it. The challenge in this regard is to reliably determine the attention of passers-by. In this article, we investigate different technical methods for estimating attention in a public display scenario by measuring physical expressive features, from which attention can be derived. In the course of an experimental setup we compare a Support Vector Machine, a neural network using a Multilayer Perceptron and a Finite State Machine and compare the results to a manual reference classification. We carve out strengths and weaknesses and identify the most feasible measuring method with regard to precision of recognition and practical application.