Editorial: “Signals to Signs” – feature extraction, recognition, and multimodal fusion

Abstract

Processing of recorded or real-time signals, feature extraction, and recognition are concepts of utmost important to an affect–aware and capable system, since they offer the opportunity to machines to benefit from modeling human behavior based on theory and interpret it based on observation. This chapter discusses feature extraction and recognition based on unimodal features in the case of speech, facial expressions and gestures, and physiological signals and elaborates on attention, fusion, dynamics, and adaptation in different multimodal cases.