An Ultra Low-Power Hardware Accelerator for Acoustic Scoring in Speech Recognition.
Hamid TabaniJosé-María ArnauJordi TubellaAntonio GonzálezPublished in: PACT (2017)
Keyphrases
- speech recognition
- speech recognition systems
- speaker independent
- speech recognizers
- ultra low power
- field programmable gate array
- acoustic models
- low cost
- language model
- speech recognizer
- hidden markov models
- low power
- automatic speech recognition
- speech signal
- noisy speech
- pattern recognition
- speech synthesis
- hardware implementation
- real time
- speech processing
- speech understanding
- noisy environments
- acoustic features
- speaker dependent
- speaker identification
- embedded systems
- speech recognition technology
- isolated word
- single chip
- speech retrieval
- face recognition
- machine learning
- neural network