Multimodal Biometric Identification System Incorporating Facial, Finger, and Iris Identifiers for Enhanced Security

Abstract

User identification is very important for all industries. Traditional methods for user authentication have included password-based, token-based, and single sign-on approaches. Also, there exists biometric authentication which depends on physiological or behavioral characteristics of the users. In the current era, unimodal biometrics are popular but its susceptibility to spoof attacks leads to the development of advanced techniques. This paper introduces a multimodal biometrics approach that will combine three different biometric modalities. These include facial, fingerprint, and iris recognition. The facial recognition approach uses FaceNet models, iris recognition involves SIFT algorithms while fingerprint recognition depends on CNNs to recognize features. Combined model can be used for user authentication. Due to the implementation of three biometric modalities, in case one fails, another approach through a security question is employed. When the question is answered correctly, it shows validity of the user.