Identifying Individuals Using Multimodal Face Recognition Techniques

A biometric system which primarily based on the cues of unimodal biometric for individual identification is not always meet the desired results. The concept of multimodal biometrics for human Identification is an emerging trend. In this paper, we present state-of-the-art novel multimodal biometric system, for face recognition, which combines the similarity scores of the unimodal modalities such as appearance based and texture based techniques of face recognition, to cater the decisive results at the level of matching score. Formally, it includes the fusion of unimodal techniques to devise the multimodal models in four possible combinations such as (a) Eigenfaces and local binary pattern (LBP), (b) Fisherfaces and LBP, (c) organics’ and augmented local binary pattern (A-LBP), and (d) Fisherfaces and A-LBP. The performance of the multimodal face recognition systems is tested on the publicly available face databases such as the AT & T-ORL and the Labeled Faces in the Wild (LFW) using a new Bray Curtis dissimilarity metric. The experimental results show a significant improvement in the performance of recognition accuracies of multimodal face recognition techniques.