Viola-Jones Algorithm Based Approach for Face Detection of African Origin People and Newborn Infants

摘要

This work dispenses to design a face detection system, especially for African origin people and newborn infants. Since many years, detection of African origin people faces is the huge problem. Using ViolaJones algorithm with some specific threshold value, the face was detected from the image with high accuracy rate. In this approach, the distance between human face and camera does not affect the detection rate. Black skin faces are hard to detect in comparison to fair skin faces because the difference in intensity of contrast between the eyes, upper cheek and nose are hard to separate by the algorithm in the black faces. Complicated background or dark/shaded background creates more complication in the detection of the black skin faces.