Hybrid Level Integration of Biometric Traits for Security Applications

In reality the security is to be provided in different levels based on the application and requirement. In attendance related applications require low level security, banking applications may need medium level security and defence applications require high level security. This paper presents a hybrid-modal (Unimodal or Multimodal) biometric system which is used to provide better security to applications based on their requirement. Based on the security level, the system uses both single evidence (unimodal) for lower level and multiple evidences (multimodal) for higher level security. The developed hybrid-modal system employs one or more biometric modalities such as face, voice and fingerprint by alleviating some of the challenges identified in fingerprint, face and voice biometrics modalities. These biometric modalities are selected as they are independent, non-intrusive and robust. Depends on the applications security level requirements like low, medium and high, the number of biometric modalities are provided as evidence to the system. The developed system is tested for 60 users. The accuracy for low level security applications using either fingerprint or face or voice the accuracy of around 94%, 93% and 82% respectively have achieved. The accuracy for medium level security applications using face & fingerprint, face & voice and voice & fingerprint are 91%, 81% and 88% respectively. Further, for high level security using all the three biometric traits the accuracy of 80% is achieved. The developed system provides promising results for all level of security applications.