Multimodal Biometric Authentication Systems: Exploring Iris and EEG Data

Authentication systems are usually classified depending on the types of characteristics they use. Biometric authentication systems rely on biological characteristics that can be unique to the individual for identification or verification purposes. Multimodal biometric authentication systems utilize multiple biometric features to improve accuracy, offset noise in collected data, or improve the trust factor of the system. As biometric data becomes more complex, many studies explored the use of deep learning and machine learning to train models that can identify or verify individuals. One of the concerns with such implementations is that they can operate as black-box solutions, which is where explainable AI (XAI) becomes favorable. This paper is an overview of the differences between unimodality and multimodality in biometric authentication systems, the identification and verification problems, deep learning and machine learning, explainable AI (XAI), and various implementations of iris and EEG data in biometric authentication systems.