Abstract
Biometric recognition is the automatic identification of individuals based on their unique behavioral or biological traits, creating a direct link to their identity that cannot be lost, shared, or duplicated. Amultimodal biometric system integrates multiple biometric traits to enhance identification accuracy and security. This research presents an optimum fusion model that incorporates iris, fingerprint, and ear biometrics, employing feature-level fusion methodology. Local binary patterns (LBP) have been used on individual biometric traits to extract unique texture features and the Bayesian optimized support vector machine feature fusion (BSFF) technique for efficient and accurate classification has been proposed. The suggested model improves the performance and robustness of biometric authentication, making it appropriate for high-security applications. Experimental studies show that the proposed method achieves an average equal error rate (EER) of 0.1478, NRMSE of 0.0983, and accuracy of 0.9343 using fivefold cross-validation.