Extremely simple fuzzy classifier for intelligent multimodal media processing

Fuzzy classifiers are known to be simple yet efficient and effective in various intelligent media processing. We introduce an extremely simple one that has been deployed in the intelligent multimodal media processing for education assessment. While this paper mainly introduces its computational model such as classification and learning (initial identification and reinforcement), this fuzzy classifier is robust against severe imbalancement of sample data distribution for learning and is consistent between fuzzy sets and probability distributions (as the normalized form of histograms). We presented an intelligent recognition of student engagement in classrooms as its application.