A Systematic Analysis of Face and Fingerprint Biometric Fusion

This paper presents various techniques used for extracting the features of face and fingerprint biometrics. More reliable recognition and performance can be achieved if both face and fingerprint modalities are fused together. The problem in multimodal biometric fusion is that the features of each modality are extracted with different algorithm which makes it difficult to fuse them together. The work presented in this paper focuses on the bimodal biometric system having Gabor filter algorithm for feature extraction as it is capable of extracting the features of both face and fingerprint biometrics. Along with this various techniques used for the fusion of two unimodal modalities have been discussed and it has been found that feature level fusion is best among all as accuracy up to 99.25% can be achieved using Gabor filter. Also the performance characteristics of all the techniques are shown on the basis of FAR (False Acceptance Rate) and FRR (False Rejection Rate) values.