Real-time Face Recognition And Its Application Based On Android System

There are so many traditional ways to authenticate someone’s id, like ID card, password. But they have their own shortcomings, they are so lacked of security and stability. With the high speed development of technology, more and more new techs are used to authenticate, these techs mainly use of our body’s biological characteristics, like fingerprint, face, iris and sound. These techs use the technology of image process and pattern recognition to identify the characters which extract from our body, so their security and stability are advanced. Nowadays, some important places are using the biological identity techs, like banks and Police.In the traditional technology of biological indentify, fingerprint are used most, and the tech are most mature. But the fingerprint has obvious shortcomings, like we must use the special equipment to extract, and the accuracy are usually not ideal and the accuracy are influenced by other factors easily. The technology of face recognition have more advantages, they are more convenient to extract faces and the faces are usually not the same, we can also extract someone’s face without his/her attention. It can’t disturb users. So, the technology of face recognition are worth of more attention because of its broad space for development.In our text, we study the technology of face recognit ion first, then we use an open-source library of Computer Vision—OpenC V, add the Adaboost and Haar-like rectangle features to design a cascading classifier, accomplish a system of Face Detect. In the process of Face Recognition, in order to solve the influence of light to the faces, we use the gray processing, geometry normalizing and so on to handle the pictures. Then we use Gabor Transfer to extract the characteristic of the pictures. As the same time, we add PCA and Gabor Transfer to reduce dimension of the characteristic. Then, we design a real-time face recognition system based on Android and accomplish it with programming, and advance the system with a function of application lock to verify the face recognition system our accomplished is really real-time with high accuracy. At last, we use the tests to verity that the algorithms chosen in our system are reliable, and they can achieve our expect. Some areas to improve are put forward in the end of the text.