Modeling Latent Discriminative Dynamic of Multi-Dimensional Affective Signals (bibtex)
by Ramirez, Geovany A., Baltrusaitis, Tadas and Morency, Louis-Philippe
Abstract:
During face-to-face communication, people continuously ex- change para-linguistic information such as their emotional state through facial expressions, posture shifts, gaze patterns and prosody. These af- fective signals are subtle and complex. In this paper, we propose to ex- plicitly model the interaction between the high level perceptual features using Latent-Dynamic Conditional Random Fields. This approach has the advantage of explicitly learning the sub-structure of the affective signals as well as the extrinsic dynamic between emotional labels. We evaluate our approach on the Audio-Visual Emotion Challenge (AVEC 2011) dataset. By using visual features easily computable using off-the- shelf sensing software (vertical and horizontal eye gaze, head tilt and smile intensity), we show that our approach based on LDCRF model outperforms previously published baselines for all four affective dimen- sions. By integrating audio features, our approach also outperforms the audio-visual baseline.
Reference:
Modeling Latent Discriminative Dynamic of Multi-Dimensional Affective Signals (Ramirez, Geovany A., Baltrusaitis, Tadas and Morency, Louis-Philippe), In Audio/Visual Emotion Challenge and Workshop (AVEC 2011), 2011.
Bibtex Entry:
@inproceedings{ramirez_modeling_2011,
	address = {Memphis, TN},
	title = {Modeling {Latent} {Discriminative} {Dynamic} of {Multi}-{Dimensional} {Affective} {Signals}},
	url = {http://ict.usc.edu/pubs/Modeling%20Latent%20Discriminative%20Dynamic%20of%20Multi-Dimensional%20Affective%20Signals.pdf},
	abstract = {During face-to-face communication, people continuously ex- change para-linguistic information such as their emotional state through facial expressions, posture shifts, gaze patterns and prosody. These af- fective signals are subtle and complex. In this paper, we propose to ex- plicitly model the interaction between the high level perceptual features using Latent-Dynamic Conditional Random Fields. This approach has the advantage of explicitly learning the sub-structure of the affective signals as well as the extrinsic dynamic between emotional labels. We evaluate our approach on the Audio-Visual Emotion Challenge (AVEC 2011) dataset. By using visual features easily computable using off-the- shelf sensing software (vertical and horizontal eye gaze, head tilt and smile intensity), we show that our approach based on LDCRF model outperforms previously published baselines for all four affective dimen- sions. By integrating audio features, our approach also outperforms the audio-visual baseline.},
	booktitle = {Audio/{Visual} {Emotion} {Challenge} and {Workshop} ({AVEC} 2011)},
	author = {Ramirez, Geovany A. and Baltrusaitis, Tadas and Morency, Louis-Philippe},
	month = oct,
	year = {2011}
}
Powered by bibtexbrowser