Cuckoo Search based Optimization for Multimodal Biometrics (Signature, Speech and Palmprint)

Biometric system based on unimodel biometrics are not able to meet the desired performance requirements for security purpose due to problems such as intra-class variations, noisy data, unacceptable error rates and low performance. Multimodal biometrics refers to the use of two or more biometrics modalities in a single identification system. The combination of different modalities is used to improve the recognition accuracy. In this proposed work, for identification based multimodal offline signature verification system, speech verification system and palmprint are combined to enhance the security and accuracy. Signature and speech are naturally produced and palmprint cannot be changed or lost. SIFT is used to extract the features of offline signature, MFCC is used to extract the features of speech and ASM is used to extract the features of palmprint. Using a technique named sum rule uses fusion at feature level in this work. To enhance the accuracy the nature based Cuckoo Search Algorithm is used to extract the best values as optimization technique. The experimental results demonstrated that the proposed multimodal biometrics system achieves a recognition accuracy of 100% with false rejection rate (FRR) of = 0. 0% & false acceptance rate (FAR) of = 0. 0%.