Biometrics is an important application of digital image processing of biometric modalities. There are several types of biometric security technologies such as face recognition, iris recognition, fingerprint matching etc. However, there are some prominent challenges in all biometrics dealing with single modality such as reliability, robustness etc. Therefore, multimodal biometrics is recommended, which involves more than one modality in the system. AADHAAR card is an example of multimodal biometrics.
Multimodal biometrics aims at increasing the reliability of biometric systems. In this research, four traits such as face, ear, iris and foot were used and worked on matching score level and decision level to get advantage of the multimodal approaches. The work was tested over self created database of 100 persons. Different classifier approaches for different modalities were applied; principle component analysis for face, eigen images for ear, hamming distance based technique for iris and modified sequential Haar transform for foot traits. Each biometric trait processed its information independently to calculate weights of individuals. The fusion scheme was applied on all possible combination of traits and the matching score was calculated. The integrated results of different biometric traits increased the recognition performance of the multimodal biometric system considerably