Audio-Visual Shared Emotion Representation for Robust Emotion Recognition on Modality Missing Using Hemi-hyperspherical Embedding and Latent Space Unification.
Seiichi HarataTakuto SakumaShohei KatoPublished in: HCI (38) (2022)
Keyphrases
- emotion recognition
- audio visual
- latent space
- multi modal
- low dimensional
- latent variables
- high dimensional
- emotional speech
- dimensionality reduction
- manifold learning
- transfer learning
- feature space
- multi stream
- visual data
- high dimensional data
- visual information
- low level
- facial expressions
- multimedia
- facial images
- data analysis
- pattern recognition
- pose estimation
- human computer interaction
- text mining
- image data
- feature selection