The Implementation Of Face Recognition Based On Hmm Attendance System

The modern information socitety puts forward higher requirements for identification. In recent years with the rapid development of computer technology and biomedical engineering, using the human body characteristics for identification is the perferred way. Because the method of face recognition technology is better than other physiological iderntification technology, it becomes a hot study issue. There are a lot of problems to slove if face recognition technology want to become a practical technology, such as face detection and the key points of face feature position, the problem of lighting and posture, large-scale face recognition problem and so on. This article focuses on face detection and the key points of face feature position, lighting, pose and the alorithm of face recognition to research.Face recognition technology included face detection and identification, this article mainly from the two aspects, Design and realized a face recognition system for attendance checking. The main research content is as following:1.For the problem of face detection speed, a certain improvement has benn made based on adaboost alorithm, using pyramid decomposition to improve the speed of face detection,After improved the speed raised obvious.2.For the problem of eyes accurate localization, an approach was presented to locate the position of pupil automatically. Firstly, it sets the eyes searching area via the distribution of eyes in the faces, and then uses image binarization and Canny Algorithm,it can extract the contour of image. Finally it can obtain the center coordinates of human eyes by ellipse fitting.3.Researched variety kind of image pre-processing technology. Using the located pupil realized the human face posture adjustment, gives a thorough research on illumination pre-processing, the experiment results show that the illumination mean and variance normalization has good effect.4.Studied the theory of face recognition based on HMM algorithm. Design and realized a face recognition system for attendance checking.5. The face recognition system are tested, the speed and accuracy of face detection, pupil localization, face recognition are tested respectly.This dissertation uses HMM algorithm realized a face recognition system.We test the system by self-built face database and standard face database, the speed of face detection can reach70frames per second, the face recognition rate can reach90%, the experimental results demonstrate that the system rate and the recognition rate has reached as expected.This system can be widely applied to security access control, identification and other security fields. Has a wide application in the future.