Efficient multimodal biometric system for human authentication using deep belief network and random forest network

Abstract

In the modernized world, the basic prerequisite of any individual is personal authentication for accessing the information. A control device which is proficient, high- tech, intelligent, efficient, very authentic and work in real time, to process all the applications including personal security to financial security is needed at this hour for all the human . Traditional method of authentication is based on password which consists of symbols or/and text values. Traditional methods of authentication process have many threats of attacks in breaking the password. The drawbacks in traditional methods can be overcome with the usage of individual characteristics usage in authentication process. This method is known as biometric authentication systems. Biometric system use human anatomy or their personality for authentication process. For eg., human face, ear, or voice, keystroke etc., The attributes used in biometric systems cannot be terminated or forgotten by any human like password as these characteristics comes along with any person. All the biometric modality has its own difficulties, merits and demerits. Hence no one modality can be used for best results. Now-a-days, every authentication system uses more than one biometric trait where one trait is fused with another one for better authentication. newline