Parallelized viterbi processor for 5, 000-word large-vocabulary real-time continuous speech recognition FPGA system.
Tsuyoshi FujinagaKazuo MiuraHiroki NoguchiHiroshi KawaguchiMasahiko YoshimotoPublished in: INTERSPEECH (2009)
Keyphrases
- continuous speech recognition
- real time
- high speed
- fpga device
- low cost
- single chip
- general purpose processors
- hardware implementation
- xilinx virtex
- low power
- data acquisition
- field programmable gate array
- digital signal
- hidden markov models
- distributed memory
- fpga hardware
- fpga implementation
- parallel architecture
- control system
- low power consumption
- parallel processing
- general purpose
- dedicated hardware
- signal processing
- co occurrence