Face detection system based on feature-based chrominance colour information

This paper presents a novel face detection system based on feature-based chrominance colour information from an image containing one face in indoor environment with non-uniform background. The face detection algorithm is based on the adapted chain code (ACC), eye detection and modified golden ratio (MGR). ACC is proposed to obtain a face boundary. MGR attempts to extract part of a face that includes eyes, eyebrows, nose and mouth, based on the detected eyes´ positions. Experimental results show that the proposed algorithm is able to detect a face of near frontal with high accuracy. The database consists of faces with and without spectacles, wearing headscarf and without wearing headscarf.