The Design And Implementation Of Face Detection System Based On AdaBoost

This topic comes from the company's project. For long-term employees on a business trip in the client of the company, the managers are hard to know their work status problem and design of attendance management system. This system is the subsystem of the company's video face recognition system, and the main function of the subsystem is face detection. The face contains a lot of visual information, is the important symbol of the mutual discrimination between the groups and memory. Through the recognition of the human face, we can get a person's information, such as gender, race, etc. The human face has the uniqueness and the non-replication, and has the non-mandatory, non-contact and concealment. Thus, the recognition of human face as the basis of identity is more appropriate than other biometric features.Through the research background of face detection, this paper introduces the significance of this topic and the current situation of the face detection development, detection methods have been expatiated, study in detail and compared several classical face detection algorithm. System used in some image processing methods were introduced, including reduce noise, gray scale, histogram equalization. The second value is discussed in this paper the principle and process of AdaBoost algorithm, make a detailed description of the Haar-like feature, integral image and other basic concepts, systematic study and description of the weak and strong classifier, and the cascade classifier. Based on AdaBoost, a face detection system is designed and realized.The system includes three modules:image preprocessing, classifier training, face detection. The test shows that the detection system can perform a fast automatic face detection and the detection rate can reach 95.63%.