Abstract A novel face detection algorithm applied in 360 degree panoramic images with complex background and strong interference colors is proposed. Firstly, color space is transformed from RGB into YCbCr combined with hue information to obtain regions of skin. Secondly, morphological opening and closing operators can filter the noise effectively. Thirdly, the geometric features of human face are introduced to get rid of false skin regions. Finally, face recogniton is extracted with the average face model. Through all procedures above,the experiments show that the proposed algorithm with high efficiency and strong adaptability offers a suitable and robust solution to human face detection in 360 degree panoramic images.
