Improving BLSTM RNN based Mandarin speech recognition using accent dependent bottleneck features.
Jiangyan YiHao NiZhengqi WenJianhua TaoPublished in: APSIPA (2016)
Keyphrases
- speech recognition
- spoken language
- speech recognition systems
- feature set
- automatic speech recognition
- image features
- multi modal
- keypoints
- extracted features
- lexical features
- feature extraction
- classification accuracy
- co occurrence
- language model
- hidden markov models
- semantic analysis
- feature space
- speech synthesis