Multimodal Biometric Verification Using Sparse Representation Based Classification

In this paper, we propose a multimodal verification system integrating face and ear based on sparse representation-based classification (SRC). The face and ear query samples are first encoded respectively to derive sparsity-based match scores, and which are then combined with the Sum-rule for verification. Our contribution is two-fold: (1) use two sparsity-based metrics and two biometric traits to demonstrate the superior performance of SRC-based verification to the well-known multimodal fusion methods using cosine similarity like likelihood ratio (LLR), support vector machine (SVM), and the Sum-rule; (2) empirically analyze the weakness and feasibility of SRC-based verification and validate the correlation between its performance and the inter-class separability of samples in dictionary. The SRC-based one-to-many comparison is not preferable to the biometrics or scenarios where inferior inter-class separability is available. This could be a selection guideline of SRC for verification applications.