by Pedro A Nogueira, Rui A Rodrigues, Eugénio Oliveira, Lennart E Nacke
Abstract:
With the popularity increase in affective computing techniques the number of emotion detection and recognition systems has risen considerably. However, despite their steady accuracy improvement, they are yet faced with application domain transferability and practical implementation issues. In this paper, we present a novel methodology for modelling individuals’ emotional states in multimedia interactive environments, while addressing the aforemen- tioned transferability and practical implementation issues. Our method relies on a two-layer classification process to classify Arousal and Valence based on four distinct physiological sensor inputs. The first classification layer uses several regression models to normalize each of the sensor inputs across participants and experimental conditions, while also correlating each input to either Arousal or Valence. The second classification layer then employs decision trees to merge the various regression outputs into one optimal Arousal/Valence classification. The presented method not only exhibits convincing accuracy ratings – 89% for Arousal and 84% for Valence – but also presents an adaptable and practical ap- proach at emotional state detection in interactive environment experiences.
Reference:
A Regression-Based Method for Lightweight Emotional State Detection in Interactive Environments (Pedro A Nogueira, Rui A Rodrigues, Eugénio Oliveira, Lennart E Nacke), In , Springer-Verlag Berlin Heidelberg, 2013.
Bibtex Entry:
@inproceedings{nogueira2013regression,
abstract = {With the popularity increase in affective computing techniques the number of emotion detection and recognition systems has risen considerably. However, despite their steady accuracy improvement, they are yet faced with application domain transferability and practical implementation issues. In this paper, we present a novel methodology for modelling individuals’ emotional states in multimedia interactive environments, while addressing the aforemen- tioned transferability and practical implementation issues. Our method relies on a two-layer classification process to classify Arousal and Valence based on four distinct physiological sensor inputs. The first classification layer uses several regression models to normalize each of the sensor inputs across participants and experimental conditions, while also correlating each input to either Arousal or Valence. The second classification layer then employs decision trees to merge the various regression outputs into one optimal Arousal/Valence classification. The presented method not only exhibits convincing accuracy ratings – 89% for Arousal and 84% for Valence - but also presents an adaptable and practical ap- proach at emotional state detection in interactive environment experiences.},
address = {Angra do Hero'{i}smo, Ac{c}ores, Portugal},
author = {Nogueira, Pedro A and Rodrigues, Rui A and Oliveira, Eug'{e}nio and Nacke, Lennart E},
journal = {Proceedings of EPIA 2013},
publisher = {Springer-Verlag Berlin Heidelberg},
title = {{A Regression-Based Method for Lightweight Emotional State Detection in Interactive Environments}},
url = {http://paginas.fe.up.pt/~niadr/PUBLICATIONS/2013/2013_EPIA.pdf},
year = {2013}
}