Remote Authentication System Based On Face Recognition

With the continuous development of the social economy, the company’s business isbigger and bigger. The amount of its internal information also becomes larger thanbefore. When the company needs to arrange an employee for a business trip，the staffcouldn’t carry all data. For security and convenience considerations, the employee whois on the business trip should visit the company’s data with remote authenticationmethod.Traditional authentication method depends on a password-based system. It onlychecks the legality of the password rather than the password holder’s permission fordata access. In most cases, the password is made up of a series of characters. so that it iseasily to be cracked. Since the biological characteristics of the human body has uniqueand not easy to steal and counterfeit, authentication method based on biometric has beengradually gaining in popularity.In this paper, a remote authentication system based on face recognition is designed,the main contributions are as followings:(1) An adaptive weighted nearest feature space analysis algorithm is proposed. It is anew extraction algorithm based on subspace learning. Feature space metric is used asdistance calculation method to obtain the between class scatter and within class scatter.A weight parameter is set to adjust their classification contribution difference.(2) A remote authentication system based on face recognition using C/S structure isdesigned. The client runs on UBIKIT6612development platform with Android system,so that the system can be easily ported to Android smartphones. The client isresponsible for the collection of human face images and encrypted it, the server side ismainly responsible for the received face image authentication and certification resultsback to the client, then the user will be able to access server-side data by the clientremote system.