Dynamic bit-width adaptation in DCT: An approach to trade off image quality and computation energy

J. Park; Jung Hwan Choi; Kaushik Roy

doi:10.1109/TVLSI.2009.2016839

Dynamic bit-width adaptation in DCT: An approach to trade off image quality and computation energy

J. Park, Jung Hwan Choi, Kaushik Roy

School of Electrical Engineering

Research output: Contribution to journal › Article › peer-review

85 Citations (Scopus)

Abstract

This paper presents a dynamic bit-width adaptation scheme for applications using discrete cosine transform (DCT). The technique can efficiently trade off image quality and computation energy. Based on sensitivity differences of 64 DCT coefficients, separate operand bit-widths are used for different frequency components to reduce computation energy. To select the appropriate operand bit-widths that achieve significant reduction of power consumption with minimum image quality degradation, we also propose a bit-width selection algorithm. The proposed variable bit precision DCT algorithm can be efficiently implemented using carry save adder trees. The reconfigurable DCT architecture can achieve power savings ranging from 36% to 75% compared to normal operation at the expense of minor image quality degradation.

Original language	English
Article number	5109472
Pages (from-to)	787-793
Number of pages	7
Journal	IEEE Transactions on Very Large Scale Integration (VLSI) Systems
Volume	18
Issue number	5
DOIs	https://doi.org/10.1109/TVLSI.2009.2016839
Publication status	Published - 2010 May

Bibliographical note

Funding Information:
Manuscript received April 03, 2008; revised October 09, 2008. First published June 23, 2009; current version published April 23, 2010. This work was supported in part by Semiconductor Research Corporation and by a Korea University Grant. A portion of this paper was presented in Design, Automation, & Test in Europe (DATE) 2006, Munich, Germany.

Keywords

Discrete cosine transform (DCT)
Dynamic bit-width
Low power design
Reconfigurable architecture

ASJC Scopus subject areas

Software
Hardware and Architecture
Electrical and Electronic Engineering

Access to Document

10.1109/TVLSI.2009.2016839

Cite this

@article{5b4da17292dd457b8d232137d54a463d,

title = "Dynamic bit-width adaptation in DCT: An approach to trade off image quality and computation energy",

abstract = "This paper presents a dynamic bit-width adaptation scheme for applications using discrete cosine transform (DCT). The technique can efficiently trade off image quality and computation energy. Based on sensitivity differences of 64 DCT coefficients, separate operand bit-widths are used for different frequency components to reduce computation energy. To select the appropriate operand bit-widths that achieve significant reduction of power consumption with minimum image quality degradation, we also propose a bit-width selection algorithm. The proposed variable bit precision DCT algorithm can be efficiently implemented using carry save adder trees. The reconfigurable DCT architecture can achieve power savings ranging from 36% to 75% compared to normal operation at the expense of minor image quality degradation.",

keywords = "Discrete cosine transform (DCT), Dynamic bit-width, Low power design, Reconfigurable architecture",

author = "J. Park and Choi, {Jung Hwan} and Kaushik Roy",

note = "Funding Information: Manuscript received April 03, 2008; revised October 09, 2008. First published June 23, 2009; current version published April 23, 2010. This work was supported in part by Semiconductor Research Corporation and by a Korea University Grant. A portion of this paper was presented in Design, Automation, & Test in Europe (DATE) 2006, Munich, Germany.",

year = "2010",

month = may,

doi = "10.1109/TVLSI.2009.2016839",

language = "English",

volume = "18",

pages = "787--793",

journal = "IEEE Transactions on Very Large Scale Integration (VLSI) Systems",

issn = "1063-8210",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "5",

}

TY - JOUR

T1 - Dynamic bit-width adaptation in DCT

T2 - An approach to trade off image quality and computation energy

AU - Park, J.

AU - Choi, Jung Hwan

AU - Roy, Kaushik

N1 - Funding Information: Manuscript received April 03, 2008; revised October 09, 2008. First published June 23, 2009; current version published April 23, 2010. This work was supported in part by Semiconductor Research Corporation and by a Korea University Grant. A portion of this paper was presented in Design, Automation, & Test in Europe (DATE) 2006, Munich, Germany.

PY - 2010/5

Y1 - 2010/5

N2 - This paper presents a dynamic bit-width adaptation scheme for applications using discrete cosine transform (DCT). The technique can efficiently trade off image quality and computation energy. Based on sensitivity differences of 64 DCT coefficients, separate operand bit-widths are used for different frequency components to reduce computation energy. To select the appropriate operand bit-widths that achieve significant reduction of power consumption with minimum image quality degradation, we also propose a bit-width selection algorithm. The proposed variable bit precision DCT algorithm can be efficiently implemented using carry save adder trees. The reconfigurable DCT architecture can achieve power savings ranging from 36% to 75% compared to normal operation at the expense of minor image quality degradation.

AB - This paper presents a dynamic bit-width adaptation scheme for applications using discrete cosine transform (DCT). The technique can efficiently trade off image quality and computation energy. Based on sensitivity differences of 64 DCT coefficients, separate operand bit-widths are used for different frequency components to reduce computation energy. To select the appropriate operand bit-widths that achieve significant reduction of power consumption with minimum image quality degradation, we also propose a bit-width selection algorithm. The proposed variable bit precision DCT algorithm can be efficiently implemented using carry save adder trees. The reconfigurable DCT architecture can achieve power savings ranging from 36% to 75% compared to normal operation at the expense of minor image quality degradation.

KW - Discrete cosine transform (DCT)

KW - Dynamic bit-width

KW - Low power design

KW - Reconfigurable architecture

UR - http://www.scopus.com/inward/record.url?scp=77951877211&partnerID=8YFLogxK

U2 - 10.1109/TVLSI.2009.2016839

DO - 10.1109/TVLSI.2009.2016839

M3 - Article

AN - SCOPUS:77951877211

SN - 1063-8210

VL - 18

SP - 787

EP - 793

JO - IEEE Transactions on Very Large Scale Integration (VLSI) Systems

JF - IEEE Transactions on Very Large Scale Integration (VLSI) Systems

IS - 5

M1 - 5109472

ER -

Dynamic bit-width adaptation in DCT: An approach to trade off image quality and computation energy

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this