A Multimodal biometric authentication system using decision theory, iris and speech recognition

This paper proposes a multimodal biometric system that combines iris recognition and speaker identification systems using the energy compaction and time frequency resolution of wavelet analysis., The uniqueness of iris pattern and the robusttness of speaker identification based on pitch period estimation complement each other in the proposed system. The paper also critically analyzes the implementation of Daubechies wavelets (Db3 and Db4) in the analysis of iris and speech samples with an endeavor to have a high success rate with optimal computational complexity.