Human face detection from images, based on skin color

Face detection is necessary in many applications, like those for face recognition, face tracking in video sequences, gender classification, biometric identification, Human Computer Interaction systems, and others. There are many approaches to face detection. The majority of them use a window-based classifier, which detects human faces by translating a window on the entire image. This detection method is not independent of illumination conditions, scale variation or pose variation. In this paper, we propose a skin color based face detection technique. First, the input image is resized and light corrected. Then, the resulted image is segmented based on skin color, using a combination of segmentation in RGB and HSV color spaces, followed by segmentation in YCbCr using the Elliptical model. The skin tone segmented image is then combined with the edges from the input image and further processed by applying the morphological operations of erosion and holes filling. In the last stage the bounding box of the detected face is displayed. The experimental results show that our method has a high accuracy in face detection from single face images. The detected faces can be of different sizes, in different poses and expressions, under unconstrained lighting conditions.