Predicting Intelligibility of Enhanced Speech Using Posteriors Derived from DNN-Based ASR System.
Kenichi AraiShoko ArakiAtsunori OgawaKeisuke KinoshitaTomohiro NakataniToshio IrinoPublished in: INTERSPEECH (2020)
Keyphrases
- speech recognition
- automatic speech recognition
- speech signal
- word error rate
- hidden markov models
- noisy environments
- speech synthesis
- language model
- conversational speech
- spontaneous speech
- speaker identification
- broadcast news
- speech retrieval
- speech recognizer
- signal to noise ratio
- pattern recognition
- handwriting recognition
- training process
- speaker diarization
- vocal tract
- audio visual
- bayesian inference
- speech corpus
- spoken words