Real time multi face detection

• Main Author: Sabado, Faith
• Other Authors: Er Meng Joo
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/61592
• Institution: Nanyang Technological University

Face detection has become the center of attraction over the years among the researchers in the computer vision field. Not only because it provides as initial platform for face recognition but it is also used in many applications such as video surveillance, responsive user interface, intelligent focus and in the recent years, in social networking sites such as Facebook. Face detection could be implemented using many existing algorithms. However existing methods use a traditional approach that would be computationally expensive. Recent advances in this field show that the work proposed by Viola and Jones provides an efficient and accurate way to detect faces that would be feasible when implemented in real time. Viola Jones provides lesser computation time, however it has its limitations. As such, in this project we would be inculcating skin color modeling as pre processing to Viola Jones to offset its false positive limitations. The whole system would consist of two parts; skin modeling – to extract skin regions and Viola Jones detector – to further verify the existence of face. The human face detection system was tested in different environment conditions and results showed a slight improvement in terms of false positive rate but the time taken to detect faces increased.