Face detection in video sequences

Advanced face detection algorithms tested on static laboratory captured images are distinguished by low error rate. Real surroundings, scenarios and scene illumination can considerably increase the error rate. These variable conditions could have a negative influence on captured video sequences. Another perspective includes fact, that human´s behavior changes at different situations which could also influence performance of face detections algorithms. The goal of the work presented in this paper is to assess the influence of these conditions on error rate. This paper presents results of Viola-Jones based face detection system tested on database trustworthily representing real surroundings and conditions. Accent is put on influence of scenarios and illumination. For this purpose dedicated closed circuit TV (CCTV) system was designed, so mentioned influences were effectively tested.