1 Chapter One Introduction

1.1 research background and significance

Access control / time and attendance is an important part of each unit property / personnel management. With the rapid development of computer technology, a variety of punch card machines, magnetic attendance machines, IC machines continue to appear, instead of the traditional manual access / attendance methods to access / attendance management into a new era. In recent years, biometrics has drawn great attention and been widely used due to its characteristics of being easy to carry, not losing, not forgetting, not being borrowed or stolen. Face recognition technology as a biometric technology, it has a direct, friendly, convenient features, easy to accept by the user, it has become the most ideal basis for authentication.

At present, the countries that have mastered more advanced face recognition technologies in the world are the United States, Germany, Canada, Israel and the like. Among them, the United States A4Vision products, FAR (false acceptance rate): 0.12%, FRR (false rejection rate): 9.79%, attitude change recognition rate: 23%, facial hair occlusion recognition rate: 50%, head occlusion recognition rate: low At 5%, low light recognition rate: less than 5%. China has also been in a leading position in this field. Among them, Prof. Dr. Su Guangda, head of the Department of Electronic Science of Tsinghua University, presided over the national "Tenth Five-Year Plan" research project - face recognition system with a recognition rate of more than 80% . Domestic, Shenzhen Kejia Technology Group KOM-V620 Face Recognition Time and Attendance System, Aerospace Golden Shield's attendance system, such as subsurface attendance rate of less than 1 second, the accuracy rate of 96% or more.

With the further development of face recognition technology, it is believed that if face recognition, feature blocking, age change, illumination change, etc. can be used to make breakthroughs in technology, more recognition and response speed are needed Face recognition technology time and attendance system products will continue to emerge.

Although face recognition has many other advantages that can not be compared with other recognitions, there are many difficulties inherent in face recognition. Face recognition is considered as the most difficult research topic in the field of biometrics or even artificial intelligence

one. The difficulty of face recognition is mainly caused by the characteristics of human face as a biological feature. Visual characteristics of human face are:

First, there is not much difference between different individuals. All human faces are similar in structure, and even human organs have similar structure and appearance. Such features are advantageous for locating human faces, but are disadvantageous for using human faces to distinguish human individuals;

Second, the face shape is very unstable, people can produce a lot of facial expressions through the face changes, and in different angles of view, the visual image of the face is also very different, in addition, face recognition by light conditions (such as daytime And night, indoor and outdoor), many face masks (such as masks, sunglasses, hair, beard, etc.), age and other factors.

In face recognition, the first type of change should be amplified as a standard to distinguish individuals, while the second type of change should be eliminated, because they can represent the same individual. The first type of change is commonly referred to as inter-class difference, while the second type is called intra-class difference. For the human face, the change within the class is often greater than the change between classes, making it difficult to distinguish between individuals using the change between classes in the case of interference within the class.

Therefore, the purpose of this research is to try to make face recognition enhance the tolerance and improve the recognition rate in postural deformation, light change and so on, and build a set of practical access control / attendance system on this basis.

The access control / time and attendance system based on face recognition technology should automate the process of data collection, data statistics and information inquiry so that employees can check in and manage personnel statistics, assess employee attendance, and can accurately reflect a certain day or a month of individuals , Departments or the entire unit attendance status for managers to provide accurate attendance data. Moreover, the data should be objective and fair, accurate and reliable in calculation, and can effectively prevent the occurrence of unpleasant phenomena such as late arrival, early departure and absenteeism. At the same time, we must also avoid the interference of human factors as far as possible, embody the openness and fairness of the attendance and attendance management system and eliminate the humanization phenomenon such as attendance and attendance. In addition, the use of face recognition based on the means to obtain relatively accurate access to personnel information, for individual safety and essential areas is essential. So as to greatly improve the management efficiency, promote the formalization, standardization and scientification of management, and then provide scientific and reliable access control / personnel attendance management tools for all units. Therefore, the outcome of the project will have a broad product prospects.