An Investigation of Partition-Based and Phonetically-Aware Acoustic Features for Continuous Emotion Prediction from Speech.
Zhaocheng HuangJulien EppsPublished in: IEEE Trans. Affect. Comput. (2020)
Keyphrases
- acoustic features
- speech signal
- speaker verification
- automatic speech recognition
- emotion recognition
- music information retrieval
- visual features
- audio features
- speech recognition
- cross correlation
- prediction accuracy
- audio visual
- audio stream
- noisy environments
- non stationary
- hidden markov models
- speaker recognition
- mel frequency cepstral coefficients
- feature set
- facial expressions
- sound source
- low level
- machine learning
- neural network