Decision fusion for frontal face verification

It has been established that the combination of a set of classifiers designed for a given pattern recognition problem may achieve higher recognition/classification rates than any of the classifiers taken individually. One of the contributing factor for the improvement is the rule applied to get a unified decision and the diversity of the classifiers. Principal Components Analysis (PCA) and Linear Discriminant Analysis (LDA) are two popular approaches in face recognition and verification. The authors will demonstrate a verification performance in which the fusion of both methods produces an improved rate compared to individual performance. Tests are carried out on FERET (Facial Recognition Technology) database using a modified protocol. A major drawback in applying LDA is that it requires a large set of individual face images sample to extract the intra-class variation. Performance is presented as the rate of verification when false acceptance rate is zero, in other words, no impostors allowed. Results using fusion of three verification experts show improvement compared with the best individual expert.