Determining speaker attributes from stress-affected speech in emergency situations with hybrid SVM-DNN architecture.
Jamil AhmadMuhammad SajjadSeungmin RhoSoon-il KwonMi Young LeeSung Wook BaikPublished in: Multim. Tools Appl. (2018)
Keyphrases
- emergency situations
- speech recognition
- speaker recognition
- audio visual
- automatic speech recognition
- support vector machine svm
- support vector
- speaker verification
- speaker identification
- automatic speech recognition systems
- speaker diarization
- support vector machine
- speaker dependent
- training process
- svm classifier
- multi class
- management system
- prosodic features
- real time
- training data
- digit recognition
- automatic transcription
- hyperplane
- svm classification
- software architecture
- classification algorithm
- knn
- neural network
- data sets
- speech signal
- classification method
- text to speech
- vocal tract
- multi modal
- partially supervised
- training set
- feature selection