A Multi-Label Learning Method for Efficient Affective Detection

Yutong Wang1, Tong Wang2, Ping Gong1, Ying Wu1, Chenfei Ye, Jie Li1, Heather Ting Ma1

  • 1Harbin Institute of Technology Shenzhen Graduate School
  • 2Harbin Institute of Technology, Shenzhen Grauate School

Details

13:25 - 14:15 | Thu 16 Feb | Ballroom D | ThRPF.6

Session: Rapid Fire Session 02: Sensor Informatics I

Abstract

Biosignals-based affective computing plays an important role in human-computing interaction. In recent years, a lot of works have been successful on emotion recognition with biological signals. However most of them are computationally expensive due to the complexity of models. In this paper, we present a multi-label learning (MLL) method to map biological signals to an affective model in real time. Multi-label learning combines multiple classifiers in a same training process by evaluating correlations between labels. To evaluate the proposed model, 25 male subjects were asked to look at dynamic images, specifically chosen to elicit emotion described by different arousal and valence levels. In terms of physiological signals, electrocardiogram (ECG) and skin conductivity (SC) signals were collected for classify. By applying the MLL method to analyze the relationships between physiological signal features and affective labels, we observed that the proposed method is effective to perform affective detection with physiological signals.