Research On Face Recognition And Identity Authentication System’s Research And Implementation Based On Android Platform

In the vast majority of high safety need systems, we need a precise way to carry out Identity Authentication. The simple way of username-password has so many flows that the system will under malicious attacks. The way of combing our system with ID authentication in a mobile devices(such as Android system)with camera has feasibility, reliability and security. In this thesis, it was used in gate-guarding system.As the number of mobile phone users increases, the amount of applications in the mobile platform would rise too, however, the malicious attacks users face would also be added, so we need to increase the security in the mobile ends. For example, when you use mobile phone to verify your information in the payment, it has easy-lose and easy-falsify problems. Face recognition is one way of biometic identification, it is secure and can not be imitated. Consequently, it will work well if we add Face Recognition in the ID Identification in mobile terminals,and in the near future the system combining Android with ID Authentication will be popularized probably.The thesis analyses relative problems about ID Authentication and Face Recognition, then introduces the FRIAS system(Based on Face Recognition Identity Authentication System). The FRIAS system firstly will get a picture in the mobile terminal, then upload the photograph to the server to process smoothing transportation in order to make it clear, lastly use histogram-cluster-analysis to achieve ID recognition. In the Android system and server we realized ID recognition, main research contents and achievements as below:Firstly, I analyzed the relative methods and processes in the face recognition. The thesis simply introduced the face detection classification got by scholars at home and aboard over a long period of time. Through researching on these basic knowledge, I got an overview on face cognition. Then the thesis explained the relative knowledge about noise suppression, and introduced the 5 ?5 binomial expression filter used in the FRIAS system.Secondly, I researched the architecture and kernel of the Android system. Mobile devices especially equipped with android system popularized rapidly and the relative applications were pilling up, so it is valuable to study deeply on it.Thirdly, I analyzed the common methods and processes in the ID Authentication. In the thesis we researched and analyzed the histogram, put forward a new way combing histogram with cluster analysis to process ID identification. Experiment proved the FRIAS system has a good identification effect and the system is stable and the response time is acceptable.Lastly, I designed and realized the FRIAS system. I introduced the FRIAS system in client end and server end, then elaborated the result of ID Authentication. The FRIAS system completed the whole ID authentication process, the correct recognition rate is 90.52%, the unexpected end rate is 4.21%, the average identification-time-cost is 45 ms.