Abstract: Multimodal biometric system verifies a person’s identity based on physiological (face, iris, fingerprint) or behavioral biometric traits (voice, signature). During this work, a brand new multimodal biometric system is developed i.e. using iris and speech. Initially, Iris and Speech recognition systems area unit developed singly by extracting their features from the Independent Component Analysis (ICA) technique for iris and from Gammatone Frequency Cepstral Coefficients (GFCCs) technique for speech. In proposed work, the speech and iris traits area unit combined along and also its performance is verified throughout authentication with the help of techniques Scale Invariant Feature Transform (SIFT) and Linear Predictive Cepstral Coefficient (LPCC). The performance evaluation of proposed method is done by Falsely Accepted Rate (FAR), Falsely Rejected Rate (FRR) and Accuracy, in MATLAB environment. The results obtained by proposed method are better than ICA and GFCC methods.

Keywords: Authentication, LPCC, SIFT, Iris, Speech.