This paper concentrates on exploiting human face informa tion for surveillance applications in a home environment. The system features real-time human face detection and facial feature identification. It is our aim to insert the results in a video-security system architec ture, where MPEG-4 coding techniques enable low bit-rate video trans mission over a home network environment. The processing contains the following essential elements: (1) skin-color segmentation, (2) histogram analysis for facial feature detection, and (3) probability-based confidence value evaluation of facial features. We have tested our sequence of pro cessing algorithms on a set of video sequences. The experimental results show that our approach offers near real-time processing speed with good detection capability, but the robustness needs further improvement.
|Title of host publication||Proceedings of the 23th symposium in information theory in the Benelux, Louvain-la-Neuve, Belgium, 29-31- 05 ,2002 m|
|Editors||B. Macq, J.J. Quisquater|
|Place of Publication||Enschede|
|Publisher||Werkgemeenschap voor Informatie- en Communicatietheorie (WIC)|
|Publication status||Published - 2002|