Multi-Stream Attention-Based BLSTM with Feature Segmentation for Speech Emotion Recognition.
Yuya ChibaTakashi NoseAkinori ItoPublished in: INTERSPEECH (2020)
Keyphrases
- multi stream
- speech emotion recognition
- audio visual speech recognition
- image segmentation
- segmentation algorithm
- audio visual
- segmentation method
- multiscale
- hidden markov models
- level set
- spatio temporal
- human computer interaction
- three dimensional
- feature vectors
- image features
- multi modal
- contextual information
- segmented regions
- high level
- data sets