FPGA Implementation of Convolutional Neural Network in Speech Recognition

QU Jia-li; HE Guang-hui; WANG Guo-xing

Chinese

您当前的位置：

首页 >

文章列表页 >

FPGA Implementation of Convolutional Neural Network in Speech Recognition

更新时间：2026-09-04

- FPGA Implementation of Convolutional Neural Network in Speech Recognition
- Vol. 35, Issue 9, Pages: 37-41(2018)
- 作者机构：
  
  上海交通大学电子信息与电气工程学院,上海,200240
- 作者简介：
- 基金信息：
- DOI：
  CLC：
- Published：2018
- 稿件说明：
移动端阅览
屈家丽, 贺光辉, 王国兴. 语音识别中卷积神经网络的FPGA实现[J]. 微电子学与计算机, 2018, 35(9): 37-41, 49.

QU Jia-li, HE Guang-hui, WANG Guo-xing. FPGA Implementation of Convolutional Neural Network in Speech Recognition[J]. 2018, 35(9): 37-41.
屈家丽, 贺光辉, 王国兴. 语音识别中卷积神经网络的FPGA实现[J]. 微电子学与计算机, 2018, 35(9): 37-41, 49. DOI：

QU Jia-li, HE Guang-hui, WANG Guo-xing. FPGA Implementation of Convolutional Neural Network in Speech Recognition[J]. 2018, 35(9): 37-41. DOI：

摘要

针对应用于语音识别中的卷积神经网络

为了提高能耗比

将网络在FPGA中进行定制化实现

并通过PCIe完成了FPGA与主机的交互.对该网络定点仿真结果表明

整体采用16位定点计算可以在保证精度的情况下有效地减少数据存储量和带宽要求.采用多种复用方式进行卷积层与全连接层的速度优化

并与流水线结构相结合

提高系统了的吞吐率

在系统150 MHz的时钟频率下达到了3 715 fps的速度.

Abstract

A convolutional neural network used in speech recognition is customized in FPGA in order to improve the energy efficiency. PCIe is used to transfer data between the FPGA and PC. A 16-bit fixed data quantization is adopted after analysis and simulation. A variety of data multiplexing methods is used for speed optimization in convolutional layer and fully-connected layer. Pipe-line helps to improve the throughput of the system. This FPGA programs with 150MHz frequency and achieves the speed of 3715fps.

关键词

Keywords

references

Lecun Y, Bengio Y, Hinton G. Deep learning[J]. Nature, 2015, 7553(521): 436-444. DOI: 436

Qiu J. Going deeper with embedded FPGA platform for convolutional neural network[C]//Acm/sigda International Symposium on Field-Programmable Gate Arrays[s.l]. 2016 : 26-35.

Gokhale V. A 240 G-ops/s mobile coprocessor for deep neural networks[C]//Computer Vision and Pattern Recognition Workshops.USA Columbus, Ohio, 2014.

Ma Y. Scalable and modularized RTL compilation of convolutional neural networks onto FPGA[C]//International Conference on Field Programmable Logic and Applications. South Africa, 2016: 1-8.

方睿, 刘加贺, 薛志辉, 等.卷积神经网络的FPGA并行加速方案设计[J].计算机工程与应用, 2015, 51(8):32-36. DOI: 32

Qian Y. Very deep convolutional neural networks for noise robust speech recognition[J]. IEEE/ACM Transactions on Audio Speech % Language Processing, 2016, 24(12): 2263-2276. DOI: 2263

Cox R V. Speech and language processing for next-millennium communications services[J]. the IEEE, 2002, 88(8): 1314-1337. DOI: 1314

Chen, Y H. Eyeriss: An energy-efficient reconfigurable accelerator for deep convolutional neural networks[J]. IEEE Journal of Solid-State Circuits, 2016, 52 (1):262-263. DOI: 262

李彦冬, 郝宗波, 雷航.卷积神经网络研究综述[J].计算机应用, 2016, 36(9): 2508-2515. DOI: 2508

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

The Implementation of Activation Function of CNN Based on FPGA Using Single Precision Floating-Point-Unit

Research of Phoneme Recognition Based on Recurrent Neural Network

A Chaos Artificial Fish Swarm Algorithm for Parameters Optimization and Application of Support Vector Machine

Speech Recognition Research on Multi-GPU Parallel Framework of Deep Neural Network

A New Method of Speech Recognition Based on Double Frame Dynamic Time Wrapping

Related Author

LI Shi-hao

YING San-cong

LI Chang-jiang

HU Yan

ZHU Wen-jing

BAI Jing

YANG Ning

ZHANG Ya-ge

Related Institution

National Key Laboratory of Fundamental Science on Synethetic Vision, School of Computer Science, Sichuan University

School of Computer Science and Technology, Wuhan University of Technology

College of Information Engineering, Taiyuan University of Technology

College of Mathematics and Information Technology, Nanjing Xiaozhuang University

School of Electronic and Information Engineering, Xi'an Jiaotong University

AI问答

Postal code：100079
Tel：（010）53879206 Email：tmw@bjxintong.com.cn
Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09082226号-64 京公网安备11010602201714号
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰