A DNN architecture with independent and shared layers for multimodal fusion reports CCC scores of 0.606, 0.534, and 0.170 for arousal, valence, and liking on the AVEC development set, outperforming early and late fusion baselines.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Multimodal Fusion with Deep Neural Networks for Audio-Video Emotion Recognition
A DNN architecture with independent and shared layers for multimodal fusion reports CCC scores of 0.606, 0.534, and 0.170 for arousal, valence, and liking on the AVEC development set, outperforming early and late fusion baselines.