Three ways to adapt a CTS recognizer to unseen reverberated speech in BUT system for the ASpIRE challenge.
Martin KarafiátFrantisek GrézlLukás BurgetIgor SzökeJan CernockýPublished in: INTERSPEECH (2015)
Keyphrases
- speech recognition
- speech recognizer
- speech corpus
- digit recognition
- speech synthesis
- speaker independent
- training set
- hidden markov models
- spoken language
- automatic speech recognition
- speaker identification
- speech signal
- recognition engine
- previously unseen
- speaker dependent
- speech recognition systems
- speech recognizers
- pattern recognition
- endpoint detection
- text to speech
- handwriting recognition
- hearing impaired
- unseen data
- broadcast news
- databases
- language acquisition
- word recognition
- dialogue system
- training examples
- recognition rate
- language model
- video sequences
- machine learning