AISHELL-4: An Open Source Dataset for Speech Enhancement, Separation, Recognition and Speaker Diarization in Conference Scenario.
Yihui FuLuyao ChengShubo LvYukai JvYuxiang KongZhuo ChenYanxin HuLei XieJian WuHui BuXin XuJun DuJingdong ChenPublished in: Interspeech (2021)
Keyphrases
- speaker diarization
- speech enhancement
- noisy environments
- object recognition
- speech recognition
- speaker identification
- speaker verification
- noise reduction
- single channel
- speech signal
- signal to noise ratio
- feature extraction
- background noise
- action recognition
- pattern recognition
- linear prediction
- sound source
- gaussian mixture model
- edge detection
- bayesian information criterion
- multiresolution