Facial Emotion Ranking Under Imbalanced Conditions

The aim of emotion recognition is to establish grounds that work for different types of emotions. However, majority of the classifiers have their base from balanced datasets. There are few works that attempts to address how to approach facial emotion recognition under imbalanced condition. This paper discusses the issues related to imbalanced data distribution problem and the common strategy to deal with imbalance datasets. We propose a model capable of handling imbalance emotion datasets well in which other typical classifiers failed to address. The model actively ranks the prototype (i.e. Prototype Ranking) of the facial expression image and adopted a derivation of support vector machines in its selection so that the problem of imbalanced data distribution can be relaxed. Then, we used an Emergent Self-Organizing Map (ESOM) to cluster the ranked features to provide clusters for unsupervised classification. This work progresses by examining the efficiency of the model in evaluating the results to show that the criterion based on prototype ranking achieves good results and performs consistently well over problem domain.