Multimodal biometric recognition using iris & fingerprint: By texture feature extraction using hybrid wavelets

Biometric systems are based on capturing human bodily features and using them for authentication. Depending on single or multiple biometric traits used for authentication, unimodal or multimodal biometric systems can be implemented. Human iris & fingerprints have unique texture pattern and this can be used for identifying the person. In this paper a multimodal biometric system based on iris and fingerprint is proposed. Texture feature extraction using Hybrid wavelets is done. Fingerprint & Iris features are extracted using multilevel decomposition of captured sample image using a new family of wavelet called Hybrid wavelet. In this paper KNN classifier used for unimodal fingerprint recognition and multi-instance iris recognition. Feature vector of iris and fingerprint are combined using decision fusion technique. The FAR-FRR analysis is performed and the results suggests that the Hybrid Wavelets are having good texture feature extraction ability and proposed system has achieved up to 77% correct classification ration.