The Research Of Face Detection Under Complex Background

Being a primary phase of face recognition,inspecting human face out from the image background is an important first step. Due to the result of face detection is usually affected by the background,brightness changing or head posture of the image and so on,which makes the process of detection more complicated.In this thesis,firstly we use the AdaBoost algorithm face detection and improve training methods; and then present a multistage detection and conformed color detection algorithm from in order to find faces in the complicated background. In the selective and conformed stage,we utilize faces'geometrical characters to roughly choose between these backup areas. we use Euler number, eyes' character further choose and discards false ones,thus identifies human face areas,finally confirms using the lip.The results of experiments show that these methods can improve the detection accuracy and decrease the detection time greatly,this system can meet the requirement of real-time detection.