Speech recognition for mixed speech and music by NMF using various cost functions and noise adaptive training methods.
Naoaki HashimotoKazumasa YamamotoSeiichi NakagawaPublished in: APSIPA (2015)
Keyphrases
- speech recognition
- noisy environments
- speech recognizer
- hidden markov models
- speech signal
- speech synthesis
- cost function
- automatic speech recognition
- speech processing
- recognition engine
- speaker identification
- noisy speech
- acoustic models
- speech recognition systems
- wall street journal corpus
- speech recognition technology
- language model
- speaker adaptation
- speech recognition errors
- spectral subtraction
- pattern recognition
- isolated word
- speaker dependent
- training process