Evaluation of human silhouette detection methods for a non-cooperative biometric system

Biometric authentication systems are becoming increasingly popular and widespread as the demand for automated people identification increases. Although, several research works focused their efforts on these type of solutions, none of the commonly available systems provide a non-cooperative approach to object identification. For this reason, they are not suitable for use in some specific situations, such as people entering the stadium. In this paper, we present an evaluation of different algorithms suitable for person detection in such environment. We focus on investigating their performance and effectiveness under unconstrained conditions, such as different lighting.