A New Approach for Multi-Biometric Fusion Based on Subjective Logic
Biometric verification systems have to address many practical requirements, such as performance, presentation attack detection (PAD), large population coverage, demographic diversity, and varied deployment environment. Traditional unimodal biometric systems do not fully meet the aforementioned requirements making them vulnerable and susceptible to different types of attacks. In response to that, modern biometric systems combine multiple biometric modalities at different fusion levels, such as sensor, feature, score and decision level. The fused score is decisive to classify an unknown user as a genuine or impostor.
In this paper, we describe a new biometric fusion framework based on Subjective Logic (SL); a type of probabilistic logic that explicitly takes uncertainty and trust into consideration. We principally evaluate our proposed fusion framework using two modalities, namely iris and fingerprint. Furthermore, the individual scores obtained from various comparators are combined at score level by applying four score fusion approaches (minimum score, maximum score, simple sum, and subjective logic) and three score normalization techniques (min-max, z-score, hyperbolic tangent). The experimental results show that the proposed score level fusion approach (subjective logic) gives the best authentication accuracy even when particular biometric classifiers give distinct comparison scores.