Comparative experiments to evaluate a voiced-unvoiced-based pre-processing approach to robust automatic speech recognition in low-SNR environments.
Hesham TolbaDouglas D. O'ShaughnessyPublished in: ICSLP (1998)
Keyphrases
- speech signal
- automatic speech recognition
- low snr
- noisy environments
- speech recognition
- preprocessing
- speaker identification
- hidden markov models
- non stationary
- word error rate
- vocal tract
- feature extraction
- speech segments
- acoustic features
- broadcast news
- recognition errors
- conversational speech
- speech retrieval
- noise model
- word recognition
- spontaneous speech
- speech corpus
- spoken words