Studying the role of pitch-adaptive spectral estimation and speaking-rate normalization in automatic speech recognition.
Syed ShahnawazuddinNagaraj AdigaHemant Kumar KathaniaGayadhar PradhanRohit SinhaPublished in: Digit. Signal Process. (2018)
Keyphrases
- linear prediction
- minimum variance
- speech signal
- spectral estimation
- automatic speech recognition
- acoustic features
- speech recognition
- hidden markov models
- noisy environments
- word error rate
- broadcast news
- conversational speech
- speech retrieval
- speech corpus
- pattern recognition
- spontaneous speech
- power spectrum
- markov random field
- feature vectors
- machine learning