New Adaptive Feature Vector Construction Procedure for Speaker Emotion Recognition Based on Wavelet Transform and Genetic Algorithm.
Alexander M. SorokaPavel E. KovaletsIgor E. KheidorovPublished in: ISNN (2016)
Keyphrases
- emotion recognition
- wavelet transform
- feature vectors
- speaker verification
- audio visual
- genetic algorithm
- feature extraction
- emotional speech
- human computer interaction
- speaker recognition
- multi modal
- image features
- multiresolution
- facial expressions
- multiscale
- neural network
- emotional state
- feature set
- mel frequency cepstral coefficients
- emotion classification
- sentiment analysis
- visual information
- similarity measure
- multimedia
- evolutionary algorithm
- speaker diarization
- machine learning