fMPE: Discriminatively Trained Features for Speech Recognition.
Daniel PoveyBrian KingsburyLidia ManguGeorge SaonHagen SoltauGeoffrey ZweigPublished in: ICASSP (1) (2005)
Keyphrases
- speech recognition
- discriminatively trained
- speech recognition systems
- hidden markov models
- object detection
- language model
- automatic speech recognition
- speech signal
- sequence classification
- speech recognizer
- speech synthesis
- discriminative training
- cepstral coefficients
- low level
- speech recognition technology
- discriminative learning
- noisy environments
- image features
- speaker identification
- feature set
- co occurrence
- feature vectors
- speaker independent
- feature space
- feature extraction
- image processing