Dual Transformation Bimodal Biometrics Based on Feature Level Fusion

The authentication of a person using bimodal biometric is robust and accurate compared to single biometric traits. In this paper we propose Dual Transformation Bimodal Biometrics Based on Feature Level Fusion (DBFLF). The Fingerprint and Face of each person is considered to identify a person. The fingerprint is preprocessed to obtain Region of Interest (ROI). The DTCWT is applied on ROI of fingerprint image to compute absolute values of high and low frequency components. The log is applied on concatenated absolute value of high and low frequency components to derive features of fingerprint. The face is preprocessed to get ROI and Harr wavelet is applied on ROI. The features of face are coefficients of approximation band. The fingerprint and face features are concatenated to derive final feature set of bimodal biometrics. The test bimodal biometric final features are compared with features of bimodal biometrics in database using Euclidian distance for matching. It is observed that the values of EER and TSR are better in the case of proposed algorithm compared to individual transformation domain techniques.