FPGA Implementation of a Pipelined Gaussian Calculation for HMM-Based Large Vocabulary Speech Recognition
FPGA Implementation of a Pipelined Gaussian Calculation for HMM-Based Large Vocabulary Speech Recognition
复制标题
基于 HMM 的大词汇量语音识别的流水线高斯计算的 FPGA 实现
DOI：
10.1155/2011/697080
复制
发表时间：
2011
影响因子：
4.3
通讯作者：
Veitch R
中科院分区：
文献类型：
--
作者：
Veitch R
A scalable large vocabulary, speaker independent speech recognition system is being developed using Hidden Markov Models (HMMs) for acoustic modeling and a Weighted Finite State Transducer (WFST) to compile sentence, word, and phoneme models. The system comprises a software backend search and an FPGA‐based Gaussian calculation which are covered here. In this paper, we present an efficient pipelined design implemented both as an embedded peripheral and as a scalable, parallel hardware accelerator. Both architectures have been implemented on an Alpha Data XRC‐5T1, reconfigurable computer housing a Virtex 5 SX95T FPGA. The core has been tested and is capable of calculating a full set of Gaussian results from 3825 acoustic models in 9.03 ms which coupled with a backend search of 5000 words has provided an accuracy of over 80%. Parallel implementations have been designed with up to 32 cores and have been successfully implemented with a clock frequency of 133 MHz.
登录
查看更多内容
用于基于分组和无线通信的低比特率分布式语音识别
DOI：
--
发表时间：
2002
期刊：
IEEE Transactions on Speech and Audio Processing
影响因子：
--
作者：
A. Bernard;A. Alwan
通讯作者：
A. Alwan
使用集成前端根据 MFCC 矢量和基频进行干净的语音重建
影响因子：
3.2
作者：
B. Milner;Xu Shao
通讯作者：
Xu Shao
基于 FPGA 软核的语音识别系统的架构设计与实现
DOI：
--
发表时间：
2006
期刊：
International Workshop System-on-Chip for Real-Time Applications
影响因子：
--
作者：
Kisun You;Hyun;Wonyong Sung
通讯作者：
Wonyong Sung
用于 5000 字词汇语音识别器的多 fpga 10 倍实时高速搜索引擎
DOI：
10.1145/1508128.1508141
发表时间：
2009
期刊：
Int. J. Reconfigurable Comput.
影响因子：
--
作者：
Edward C. Lin;Rob A. Rutenbar
通讯作者：
Rob A. Rutenbar