Melnikoff, Stephen Jonathan and Quigley, Steven Francis and Russell, Martin (2002) Performing speech recognition on multiple parallel files using continuous hidden Markov models on an FPGA. In: Field-Programmable Technology, 2002. (FPT). Proceedings. IEEE, pp. 399-402.
URL of Published Version: http://ieeexplore.ieee.org/iel5/8456/26638/01188720.pdf?isnumber=26638&prod=STD&arnumber=1188720&arnumber=1188720&arSt=+399&ared=+402&arAuthor=Melnikoff%2C+S.J.%3B+Quigley%2C+S.F.%3B+Russell%2C+M.J.
Speech recognition is a computationally demanding task, Particularly the stages which use Viterbi decoding for converting pre-processed speech data into words or subword unit, and the associated observation probability calculations, which employ multivariate Gaussian distributions; so any device that can reduce the load on, for example, a PC's processor, is advantageous. Hence we present two implementations of a speech recognition system incorporating an FPGA, employing continuous hidden Markov models (HMMs), and capable of processing three speech files simultaneously. The first uses monophones, and can perform recognition 250 times real time (in terms of average time per observation), as well as outperforming its software equivalent. The second uses biphones and triphones, reducing the speedup to 13 times real time.
|Type of Work:||Book Section|
|School/Faculty:||Schools (1998 to 2008) > School of Engineering|
|Department:||Electronic, Electrical and Computer Engineering|
©2002 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
|Subjects:||TK Electrical engineering. Electronics Nuclear engineering|
QA75 Electronic computers. Computer science
|Institution:||University of Birmingham|
Repository Staff Only: item control page