A Boosting-Based Decision Fusion Method for Learning from Large, Imbalanced Face Data Set

The acquisition of face images is usually limited due to policy and economy considerations, and hence the number of training examples of each subject varies greatly. The problem of face recognition with imbalanced training data has drawn attention of researchers and it is desirable to understand in what circumstances imbalanced data set affects the learning outcomes, and robust methods are needed to maximize the information embedded in the training data set without relying much on user introduced bias. In this article, we study the effects of uneven number of training images for automatic face recognition and proposed a boosting-based decision fusion method that suppresses the face recognition errors by training an ensemble with subsets of examples. By recovering the balance among classes in the subsets, our proposed multiBoost.imb method circumvents the class skewness and demonstrates improved performance. Experiments are conducted with four popular face data sets and two synthetic data sets. The results of our method exhibits superior performance in high imbalanced scenarios compared to AdaBoost.M1, SAMME, RUSboost, SMOTEboost, SAMME with SMOTE sampling and SAMME with random undersampling. Another advantage that comes with using subsets of examples is the significant gain in efficiency.