Robust Component-based Face Detection Using Color Feature

Face detection is an important topic in many applications. Variation of illumination and pose in addition to existence occlusion and orientation decrease the algorithm performance. These factors change global facial appearance. Using component-based methods, we can overcome the referred problems. Because of face components are less affected by these factors. In this paper, we propose a novel component-based face detection approach based on color features. At first, the skin color regions are segmented in transformed color spaces. Then the eyes and mouth are localized as necessary components in face candidate regions. Finally, face or faces are detected by verifying the geometric relations between the three facial components. Our proposed method is even able to localize occluded facial features. Furthermore we can detect not only single face but also multiple faces in an image.