An Efficient FPGA Implementation of Gaussian Mixture Models-Based Classifier Using Distributed Arithmetic

Minghua Shi, Amine Bermak, S. Chandrasekaran, A. Amira

Research output: Chapter in Book/Report/Conference proceedingConference contribution

15 Citations (Scopus)

Abstract

Gaussian Mixture Models (GMM)-based classifiers have shown increased attention in many pattern recognition applications. Improved performances have been demonstrated in many applications but using such classifiers can require large storage and complex processing units due to exponential calculations and large number of coefficients involved. This poses a serious problem for portable real-time pattern recognition applications. In this paper, first the performance of GMM and its hardware complexity are analyzed and compared with a number of benchmark algorithms. Next, an efficient digital hardware implementation based on Distributed Arithmetic (DA) is proposed. A novel exponential calculation circuit based on linear piecewise approximation is also developed to reduce hardware complexity. Implementation is carried out on the Celoxica-RC1000 board equipped with the Virtex-E FPGA. Maximum optimization has been achieved by means of manual placement and routing in order to achieve a compact core footprint. A detailed evaluation of the performance metrics of the GMM core is also presented.

Original languageEnglish
Title of host publicationICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems
Pages1276-1279
Number of pages4
DOIs
Publication statusPublished - 2006
Externally publishedYes
EventICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems - Nice, France
Duration: 10 Dec 200613 Dec 2006

Other

OtherICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems
CountryFrance
CityNice
Period10/12/0613/12/06

Fingerprint

Field programmable gate arrays (FPGA)
Classifiers
Hardware
Pattern recognition
Networks (circuits)
Processing

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Shi, M., Bermak, A., Chandrasekaran, S., & Amira, A. (2006). An Efficient FPGA Implementation of Gaussian Mixture Models-Based Classifier Using Distributed Arithmetic. In ICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems (pp. 1276-1279). [4263607] https://doi.org/10.1109/ICECS.2006.379695

An Efficient FPGA Implementation of Gaussian Mixture Models-Based Classifier Using Distributed Arithmetic. / Shi, Minghua; Bermak, Amine; Chandrasekaran, S.; Amira, A.

ICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems. 2006. p. 1276-1279 4263607.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Shi, M, Bermak, A, Chandrasekaran, S & Amira, A 2006, An Efficient FPGA Implementation of Gaussian Mixture Models-Based Classifier Using Distributed Arithmetic. in ICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems., 4263607, pp. 1276-1279, ICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems, Nice, France, 10/12/06. https://doi.org/10.1109/ICECS.2006.379695
Shi M, Bermak A, Chandrasekaran S, Amira A. An Efficient FPGA Implementation of Gaussian Mixture Models-Based Classifier Using Distributed Arithmetic. In ICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems. 2006. p. 1276-1279. 4263607 https://doi.org/10.1109/ICECS.2006.379695
Shi, Minghua ; Bermak, Amine ; Chandrasekaran, S. ; Amira, A. / An Efficient FPGA Implementation of Gaussian Mixture Models-Based Classifier Using Distributed Arithmetic. ICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems. 2006. pp. 1276-1279
@inproceedings{36b7a43750db42638d2455c9570eac51,
title = "An Efficient FPGA Implementation of Gaussian Mixture Models-Based Classifier Using Distributed Arithmetic",
abstract = "Gaussian Mixture Models (GMM)-based classifiers have shown increased attention in many pattern recognition applications. Improved performances have been demonstrated in many applications but using such classifiers can require large storage and complex processing units due to exponential calculations and large number of coefficients involved. This poses a serious problem for portable real-time pattern recognition applications. In this paper, first the performance of GMM and its hardware complexity are analyzed and compared with a number of benchmark algorithms. Next, an efficient digital hardware implementation based on Distributed Arithmetic (DA) is proposed. A novel exponential calculation circuit based on linear piecewise approximation is also developed to reduce hardware complexity. Implementation is carried out on the Celoxica-RC1000 board equipped with the Virtex-E FPGA. Maximum optimization has been achieved by means of manual placement and routing in order to achieve a compact core footprint. A detailed evaluation of the performance metrics of the GMM core is also presented.",
author = "Minghua Shi and Amine Bermak and S. Chandrasekaran and A. Amira",
year = "2006",
doi = "10.1109/ICECS.2006.379695",
language = "English",
isbn = "1424403952",
pages = "1276--1279",
booktitle = "ICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems",

}

TY - GEN

T1 - An Efficient FPGA Implementation of Gaussian Mixture Models-Based Classifier Using Distributed Arithmetic

AU - Shi, Minghua

AU - Bermak, Amine

AU - Chandrasekaran, S.

AU - Amira, A.

PY - 2006

Y1 - 2006

N2 - Gaussian Mixture Models (GMM)-based classifiers have shown increased attention in many pattern recognition applications. Improved performances have been demonstrated in many applications but using such classifiers can require large storage and complex processing units due to exponential calculations and large number of coefficients involved. This poses a serious problem for portable real-time pattern recognition applications. In this paper, first the performance of GMM and its hardware complexity are analyzed and compared with a number of benchmark algorithms. Next, an efficient digital hardware implementation based on Distributed Arithmetic (DA) is proposed. A novel exponential calculation circuit based on linear piecewise approximation is also developed to reduce hardware complexity. Implementation is carried out on the Celoxica-RC1000 board equipped with the Virtex-E FPGA. Maximum optimization has been achieved by means of manual placement and routing in order to achieve a compact core footprint. A detailed evaluation of the performance metrics of the GMM core is also presented.

AB - Gaussian Mixture Models (GMM)-based classifiers have shown increased attention in many pattern recognition applications. Improved performances have been demonstrated in many applications but using such classifiers can require large storage and complex processing units due to exponential calculations and large number of coefficients involved. This poses a serious problem for portable real-time pattern recognition applications. In this paper, first the performance of GMM and its hardware complexity are analyzed and compared with a number of benchmark algorithms. Next, an efficient digital hardware implementation based on Distributed Arithmetic (DA) is proposed. A novel exponential calculation circuit based on linear piecewise approximation is also developed to reduce hardware complexity. Implementation is carried out on the Celoxica-RC1000 board equipped with the Virtex-E FPGA. Maximum optimization has been achieved by means of manual placement and routing in order to achieve a compact core footprint. A detailed evaluation of the performance metrics of the GMM core is also presented.

UR - http://www.scopus.com/inward/record.url?scp=47349089959&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=47349089959&partnerID=8YFLogxK

U2 - 10.1109/ICECS.2006.379695

DO - 10.1109/ICECS.2006.379695

M3 - Conference contribution

SN - 1424403952

SN - 9781424403950

SP - 1276

EP - 1279

BT - ICECS 2006 - 13th IEEE International Conference on Electronics, Circuits and Systems

ER -