Stimulus synthesis for efficient evaluation and refinement of perceptual image quality metrics

Zhou Wang, Eero Simoncelli

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

Abstract

We propose a methodology for comparing and refining perceptual image quality metrics based on synthetic images that are optimized to best differentiate two candidate quality metrics. We start from an initial distorted image and iteratively search for the best/worst images in terms of one metric while constraining the value of the other to remain fixed. We then repeat this, reversing the roles of the two metrics. Subjective test on the quality of pairs of these images generated at different initial distortion levels provides a strong indication of the relative strength and weaknesses of the metrics being compared. This methodology also provides an efficient way to further refine the definition of an image quality metric.

Original languageEnglish (US)
Title of host publicationProceedings of SPIE - The International Society for Optical Engineering
EditorsB.E. Rogowitz, T.N. Pappas
Pages99-108
Number of pages10
Volume5292
DOIs
StatePublished - 2004
EventHuman Vision and Electronic Imaging IX - San Jose, CA, United States
Duration: Jan 19 2004Jan 21 2004

Other

OtherHuman Vision and Electronic Imaging IX
CountryUnited States
CitySan Jose, CA
Period1/19/041/21/04

Fingerprint

stimuli
Image quality
evaluation
synthesis
Refining
methodology
reversing
refining
indication

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Condensed Matter Physics

Cite this

Wang, Z., & Simoncelli, E. (2004). Stimulus synthesis for efficient evaluation and refinement of perceptual image quality metrics. In B. E. Rogowitz, & T. N. Pappas (Eds.), Proceedings of SPIE - The International Society for Optical Engineering (Vol. 5292, pp. 99-108) https://doi.org/10.1117/12.537129

Stimulus synthesis for efficient evaluation and refinement of perceptual image quality metrics. / Wang, Zhou; Simoncelli, Eero.

Proceedings of SPIE - The International Society for Optical Engineering. ed. / B.E. Rogowitz; T.N. Pappas. Vol. 5292 2004. p. 99-108.

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

Wang, Z & Simoncelli, E 2004, Stimulus synthesis for efficient evaluation and refinement of perceptual image quality metrics. in BE Rogowitz & TN Pappas (eds), Proceedings of SPIE - The International Society for Optical Engineering. vol. 5292, pp. 99-108, Human Vision and Electronic Imaging IX, San Jose, CA, United States, 1/19/04. https://doi.org/10.1117/12.537129
Wang Z, Simoncelli E. Stimulus synthesis for efficient evaluation and refinement of perceptual image quality metrics. In Rogowitz BE, Pappas TN, editors, Proceedings of SPIE - The International Society for Optical Engineering. Vol. 5292. 2004. p. 99-108 https://doi.org/10.1117/12.537129
Wang, Zhou ; Simoncelli, Eero. / Stimulus synthesis for efficient evaluation and refinement of perceptual image quality metrics. Proceedings of SPIE - The International Society for Optical Engineering. editor / B.E. Rogowitz ; T.N. Pappas. Vol. 5292 2004. pp. 99-108
@inproceedings{c347daa238874783a0aa8d827065b1b9,
title = "Stimulus synthesis for efficient evaluation and refinement of perceptual image quality metrics",
abstract = "We propose a methodology for comparing and refining perceptual image quality metrics based on synthetic images that are optimized to best differentiate two candidate quality metrics. We start from an initial distorted image and iteratively search for the best/worst images in terms of one metric while constraining the value of the other to remain fixed. We then repeat this, reversing the roles of the two metrics. Subjective test on the quality of pairs of these images generated at different initial distortion levels provides a strong indication of the relative strength and weaknesses of the metrics being compared. This methodology also provides an efficient way to further refine the definition of an image quality metric.",
author = "Zhou Wang and Eero Simoncelli",
year = "2004",
doi = "10.1117/12.537129",
language = "English (US)",
volume = "5292",
pages = "99--108",
editor = "B.E. Rogowitz and T.N. Pappas",
booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",

}

TY - GEN

T1 - Stimulus synthesis for efficient evaluation and refinement of perceptual image quality metrics

AU - Wang, Zhou

AU - Simoncelli, Eero

PY - 2004

Y1 - 2004

N2 - We propose a methodology for comparing and refining perceptual image quality metrics based on synthetic images that are optimized to best differentiate two candidate quality metrics. We start from an initial distorted image and iteratively search for the best/worst images in terms of one metric while constraining the value of the other to remain fixed. We then repeat this, reversing the roles of the two metrics. Subjective test on the quality of pairs of these images generated at different initial distortion levels provides a strong indication of the relative strength and weaknesses of the metrics being compared. This methodology also provides an efficient way to further refine the definition of an image quality metric.

AB - We propose a methodology for comparing and refining perceptual image quality metrics based on synthetic images that are optimized to best differentiate two candidate quality metrics. We start from an initial distorted image and iteratively search for the best/worst images in terms of one metric while constraining the value of the other to remain fixed. We then repeat this, reversing the roles of the two metrics. Subjective test on the quality of pairs of these images generated at different initial distortion levels provides a strong indication of the relative strength and weaknesses of the metrics being compared. This methodology also provides an efficient way to further refine the definition of an image quality metric.

UR - http://www.scopus.com/inward/record.url?scp=8844268312&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=8844268312&partnerID=8YFLogxK

U2 - 10.1117/12.537129

DO - 10.1117/12.537129

M3 - Conference contribution

VL - 5292

SP - 99

EP - 108

BT - Proceedings of SPIE - The International Society for Optical Engineering

A2 - Rogowitz, B.E.

A2 - Pappas, T.N.

ER -