Comparison of face recognition algorithms for human-robot interactions

Abstract

Face recognition is a cornerstone of many robotic systems in which a robot has to identify and interact with a human being. Choosing a face recognition algorithm arbitrarily may not yield the best results for a researcher and may produce undermined results. In this paper we compare three widely used algorithms in terms of speed and accuracy. Such data can be very useful in choosing an algorithm for a particular task. The algorithms were applied to 36 different situations, and the results indicate the strengths, advantage and limitations of each of the three recognition methods in a certain setting.