Separate-to-Recognize: Joint Multi-target Speech Separation and Speech Recognition for Speaker-attributed ASR.
Yuxiao LinZhihao DuShiliang ZhangFan YuZhou ZhaoFei WuPublished in: ISCSLP (2022)
Keyphrases
- speech recognition
- multi target
- automatic speech recognition
- speech signal
- multi target tracking
- multi sensor
- word error rate
- hidden markov models
- speech synthesis
- speech processing
- language model
- speaker identification
- speaker dependent
- pattern recognition
- speech recognizer
- speech recognition systems
- noisy environments
- multi camera
- broadcast news
- speech recognition technology
- visual tracking
- data association
- speaker independent
- multiple targets
- isolated word
- speech recognizers
- speaker diarization
- speaker recognition
- conversational speech
- speech retrieval
- acoustic models
- infrared
- vocal tract
- sound source
- probabilistic model
- acoustic features
- speaker adaptation
- word recognition
- speaker verification
- image processing
- cepstral coefficients
- spontaneous speech
- multi modal