Robust frontal face detection in complex environment

We constructed a simple and fast system to detect frontal human faces in complex environment. There are two main contributions of our work: 1) we use a fast image segmentation method based on connected components labelling to select candidate face areas; 2) we propose a positive-negative attractor template to examine face areas. A valley detector is used to search the valley-like points of eyes and mouths. We test the system on images in complex environment and with confusing objects. The experiment shows a robust detection result with few false detected faces.