Mixed Precision DNN Quantization for Overlapped Speech Separation and Recognition.
Junhao XuJianwei YuXunying LiuHelen MengPublished in: ICASSP (2022)
Keyphrases
- recognition engine
- recognition accuracy
- recognition rate
- speech recognition
- speech corpus
- object recognition
- continuous speech recognition
- high precision
- spoken words
- recognition algorithm
- noisy environments
- automatic speech recognition systems
- recognition process
- automatic recognition
- automatic speech recognition
- text recognition
- precision and recall
- digit recognition
- handwritten characters
- speaker identification
- spoken language
- word recognition
- speech signal
- action recognition
- facial gestures
- pattern recognition
- anti noise
- feature extraction
- broadcast news
- computer vision