Research And Implementation Of Remote Face Verification System Based On Android Platform

With the rapid development of mobile internet, the security of network users’information becomes increasingly prominent. Compared with traditional identification technology, face recognition is more stable, reliable, and effective, so it’s more suitable for the security field of mobile internet. In recent years, however, face recognition technology is more frequently applied to social contact and recreational products. Researches on face verification are comparatively less. This paper combines face recognition technology, mobile internet and smart phone together, and realizes remote face verification system based on Android platform. The main points are as follows:This paper introduces basic frameworks of face identification and face verification of face recognition technology. Then it reviews image illumination processing approaches and facial features extraction approaches during current face recognition process. Based on the analysis of features of mobile internet, it proposes a solution and its realizing process of face verification system which is applied to smart phone-server under mobile environment, and it analyzes the structural features of the system.Facial images in the system, which are used to verify, are acquired through smart phone’s camera. The images mainly include illumination variation, expression changes and covering of the glasses, of which illumination variation influences the result most. In order to process face verification under unsymmetrical illumination better, this paper studies two facial image illumination processing approaches and three facial features extraction approaches in different combinations. After the experimental analysis, this paper proposes a new combination of DCT(discrete cosine transform) with Fisherface, whose equal error rates on the Extended Yale B database and AR database are1.04%and6.18%respectively. It realizes remote face verification system based on Android platform, which can be applied to aspects such as users’secure remote login. The mobile phone mainly completes face detection, image preprocessing, feature extraction and transmission of verification information. In order to ensure the transmission security of users’ verification information, the mobile phone uses AES encryption algorithm to encrypt the verification information. The server mainly completes reception and deciphering of verification information and transmits the results of feature similarity judgement to the mobile phone. After the realization of the system, the author set up a small-sized face database to test the practicability of the system.