Title
A 40-NM 54-MW 3×-real-time VLSI processor for 60-kWord continuous speech recognition
Abstract
This paper describes a low-power VLSI chip for speaker-independent 60-kWord continuous speech recognition based on a context-dependent Hidden Markov Model (HMM). We implement parallel and pipelined architecture for GMM computation and Viterbi processing. It includes a 8-path Viterbi transition architecture to maximize the processing speed and adopts tri-gram language model to improve the recognition accuracy. A two-level cache architecture is implemented for the demo system. The test chip, fabricated in 40 nm CMOS technology, occupies 1.77 mm × 2.18 mm containing 2.98 M transistors for logic and 4.29 Mbit on-chip memory. The measured results show that our implementation achieves 25% required frequency reduction (62.5 MHz) and 26% power consumption reduction (54.8 mW) for 60 k-Word real-time continuous speech recognition compared to the previous work. This chip can maximally process 3.02× and 2.25× times faster than real-time at 200 MHz using the bigram and trigram language models, respectively.
Year
DOI
Venue
2013
10.1109/SiPS.2013.6674496
SiPS
Keywords
Field
DocType
cmos integrated circuits,real-time vlsi processor,speech recognition,microprocessor chips,3×,frequency 200 mhz,pipelined architecture,parallel architectures,size 40 nm,hmm,low-power vlsi chip,viterbi transition architecture,tri-gram language,low-power electronics,40 nm vlsi,cmos technology,viterbi processing,bigram language,vlsi,power 54 mw,cache architecture,power 54.8 mw,frequency reduction,continuous speech recognition,hidden markov models,parallel architecture,gmm computation,large vocabulary continuous speech recognition (lvscr),real-time systems,hidden markov model,on-chip memory,power consumption reduction,low power electronics,real time systems
Computer science,Parallel computing,Cache-only memory architecture,Chip,CMOS,Real-time computing,Speech recognition,Bigram,Hidden Markov model,Very-large-scale integration,Viterbi algorithm,Language model
Conference
ISSN
Citations 
PageRank 
2162-3562
3
0.44
References 
Authors
8
6
Name
Order
Citations
PageRank
Guangji He1183.02
Yuki Miyamoto230.44
Kumpei Matsuda330.78
Shintaro Izumi48231.56
Hiroshi Kawaguchi53721.08
Masahiko Yoshimoto630.44