Emotion assessing using valence-arousal evaluation based on peripheral physiological signals and support vector machine

Emotion recognition becomes an investigated topic in affective computing for several applications. The presented paper aims to recognize human emotions using peripheral physiological signals as well as electrocardiogram (ECG), galvanic skin response (GSR), Skin Temperature (Temp) and respiration volume (RV). To achieve this purpose, we develop our work with the multimodal database MAHNOB-HCI. The emotional responses of twenty four participants to twenty affective stimuli videos are classified into two precise areas in valence-arousal emotional space. Using the support vector machine (SVM) as a classifier, the results, over the mentioned signals, show that the ECG and RV signals are the most relevant for emotion recognition issue. Moreover, our obtained accuracies are promising compared to related work.