PFV-ATM System Realization Based On Face And Voice

Biometrics is an emerging technology, which has unique advantages over traditional authentication technologies and has wide applications such as in financial, public security areas and in everyday life. This paper focuses on realization of PFV-ATM system based on face and voice, which also integrated with traditional information security. The main content is as follows:Firstly, Aiming at the actual application environment,we make many improvements in the algorithms of speaker and face verification,in order to get the same effect as in the lab.As to the voice,we first apply the technologies of noise and activity detection,after which we extract the feature, then we broaden the dimensions of it. Last we use the spliting method to establish the initial speaker modal. As to the face, we put forward a face detection algorithm based on complexion background modal,and apply it in our system. Through the experiments,we found that the improved voice and face recognition algorithms had an strong adaptive ablity to the changable enviroment, and gained a higher verification accuracy than former algorithms.Secondly, we adopted Client/Server modal to decompose the whole system and constructed it from 3 aspects: Client,Server,Manager.Client is mainly responsible for good GUI(Graphics User Interface) to let users collect their face and voice.Also, the client need to match the biometric file with the feature modal tranferred from server and show the vericifation result on the display; Server is mainly used to deposit the training files including face and voice feature modal.Manager is mainly for manager to upadte the users' database,including inquirying,modifying,registering and logoff.Thereinto them, registering process consisists of information collecting,feature extraction,after which the user's information is saved in the IC card and feature modal is saved in the server database.Finally, From the usual attacks in the E-commerce and ATM systems ,concerning of our project, we put forward a security strategy based on traditional information security technology and biometrics recognition and realize it.