Prosodic, Spectral and Voice Quality Feature Selection Using a Long-Term Stopping Criterion for Audio-Based Emotion Recognition.
Markus KächeleDimitri ZharkovSascha MeudtFriedhelm SchwenkerPublished in: ICPR (2014)
Keyphrases
- emotion recognition
- feature selection
- stopping criterion
- audio visual
- emotional speech
- human computer interaction
- speaker verification
- facial expressions
- prosodic features
- sentiment analysis
- information fusion
- emotion classification
- text classification
- high quality
- facial images
- feature extraction
- convergence rate
- multi modal
- machine learning
- cost function
- multiscale
- emotional state
- facial features
- computational intelligence
- input image