A computational efficient real time noise robust speech recognition based on improved spectral subtraction method.
Bojan KotnikZdravko KacicBogomir HorvatPublished in: INTERSPEECH (2001)
Keyphrases
- spectral subtraction
- speech recognition
- hybrid method
- signal to noise ratio
- noisy environments
- power spectrum
- support vector machine
- hidden markov models
- language model
- noisy speech
- speech recognizer
- information retrieval
- gaussian distribution
- noise reduction
- error rate
- pattern recognition
- speaker identification
- machine learning
- neural network