Emotion investigation based on biosignals

Abstract

In this paper a physiological signal-based emotion recognition approach is presented. The input biosignals are electromyogram, electrocardiogram, skin conductivity and respiration change. The feature vector is extracted from each signal type by using the same technique based on wavelets and TESPAR DZ method. A Support Vector Machine (SVM) classifier was employed to distinguish among four emotional states: joy, anger, sadness and pleasure. The database employed in our experiments is the AuBT corpus.