Conv-Transformer Transducer: Low Latency, Low Frame Rate, Streamable End-to-End Speech Recognition.
Wenyong HuangWenchao HuYu Ting YeungXiao ChenPublished in: INTERSPEECH (2020)
Keyphrases
- speech recognition
- end to end
- low latency
- low frame rate
- frame rate
- high speed
- high bandwidth
- low resolution
- gait recognition
- high resolution
- hidden markov models
- high throughput
- language model
- pattern recognition
- highly efficient
- real time
- stream processing
- super resolution
- virtual machine
- multipath
- wireless sensor networks
- neural network
- data mining
- ad hoc networks
- video sequences
- data processing
- computer vision
- low complexity
- information retrieval
- image quality