Perceptual image quality assessment using a normalized Laplacian pyramid

Valero Laparra, Johannes Balle, Alexander Berardino, Eero Simoncelli

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

Abstract

We present an image quality metric based on the transformations associated with the early visual system: local luminance subtraction and local gain control. Images are decomposed using a Laplacian pyramid, which subtracts a local estimate of the mean luminance at multiple scales. Each pyramid coefficient is then divided by a local estimate of amplitude (weighted sum of absolute values of neighbors), where the weights are optimized for prediction of amplitude using (undistorted) images from a separate database. We define the quality of a distorted image, relative to its undistorted original, as the root mean squared error in this "normalized Laplacian " domain. We show that both luminance subtraction and amplitude division stages lead to significant reductions in redundancy relative to the original image pixels. We also show that the resulting quality metric provides a better account of human perceptual judgements than either MS-SSIM or a recently-published gain-control metric based on oriented filters.

Original languageEnglish (US)
Title of host publicationHuman Vision and Electronic Imaging 2016, HVEI 2016
EditorsThrasyvoulos N. Pappas, Huib de Ridder, Bernice E. Rogowitz
PublisherSociety for Imaging Science and Technology
Pages43-48
Number of pages6
ISBN (Electronic)9781510827943
DOIs
StatePublished - Jan 1 2016
EventHuman Vision and Electronic Imaging 2016, HVEI 2016 - San Francisco, United States
Duration: Feb 14 2016Feb 18 2016

Publication series

NameHuman Vision and Electronic Imaging 2016, HVEI 2016

Conference

ConferenceHuman Vision and Electronic Imaging 2016, HVEI 2016
CountryUnited States
CitySan Francisco
Period2/14/162/18/16

Fingerprint

pyramids
Image quality
Luminance
Gain control
luminance
subtraction
Redundancy
Pixels
redundancy
estimates
division
pixels
filters
coefficients
predictions

ASJC Scopus subject areas

  • Atomic and Molecular Physics, and Optics
  • Electrical and Electronic Engineering
  • Computer Science Applications

Cite this

Laparra, V., Balle, J., Berardino, A., & Simoncelli, E. (2016). Perceptual image quality assessment using a normalized Laplacian pyramid. In T. N. Pappas, H. de Ridder, & B. E. Rogowitz (Eds.), Human Vision and Electronic Imaging 2016, HVEI 2016 (pp. 43-48). (Human Vision and Electronic Imaging 2016, HVEI 2016). Society for Imaging Science and Technology. https://doi.org/10.2352/ISSN.2470-1173.2016.16HVEI-103

Perceptual image quality assessment using a normalized Laplacian pyramid. / Laparra, Valero; Balle, Johannes; Berardino, Alexander; Simoncelli, Eero.

Human Vision and Electronic Imaging 2016, HVEI 2016. ed. / Thrasyvoulos N. Pappas; Huib de Ridder; Bernice E. Rogowitz. Society for Imaging Science and Technology, 2016. p. 43-48 (Human Vision and Electronic Imaging 2016, HVEI 2016).

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

Laparra, V, Balle, J, Berardino, A & Simoncelli, E 2016, Perceptual image quality assessment using a normalized Laplacian pyramid. in TN Pappas, H de Ridder & BE Rogowitz (eds), Human Vision and Electronic Imaging 2016, HVEI 2016. Human Vision and Electronic Imaging 2016, HVEI 2016, Society for Imaging Science and Technology, pp. 43-48, Human Vision and Electronic Imaging 2016, HVEI 2016, San Francisco, United States, 2/14/16. https://doi.org/10.2352/ISSN.2470-1173.2016.16HVEI-103
Laparra V, Balle J, Berardino A, Simoncelli E. Perceptual image quality assessment using a normalized Laplacian pyramid. In Pappas TN, de Ridder H, Rogowitz BE, editors, Human Vision and Electronic Imaging 2016, HVEI 2016. Society for Imaging Science and Technology. 2016. p. 43-48. (Human Vision and Electronic Imaging 2016, HVEI 2016). https://doi.org/10.2352/ISSN.2470-1173.2016.16HVEI-103
Laparra, Valero ; Balle, Johannes ; Berardino, Alexander ; Simoncelli, Eero. / Perceptual image quality assessment using a normalized Laplacian pyramid. Human Vision and Electronic Imaging 2016, HVEI 2016. editor / Thrasyvoulos N. Pappas ; Huib de Ridder ; Bernice E. Rogowitz. Society for Imaging Science and Technology, 2016. pp. 43-48 (Human Vision and Electronic Imaging 2016, HVEI 2016).
@inproceedings{e093d77d718b4e2d94fa529683840244,
title = "Perceptual image quality assessment using a normalized Laplacian pyramid",
abstract = "We present an image quality metric based on the transformations associated with the early visual system: local luminance subtraction and local gain control. Images are decomposed using a Laplacian pyramid, which subtracts a local estimate of the mean luminance at multiple scales. Each pyramid coefficient is then divided by a local estimate of amplitude (weighted sum of absolute values of neighbors), where the weights are optimized for prediction of amplitude using (undistorted) images from a separate database. We define the quality of a distorted image, relative to its undistorted original, as the root mean squared error in this {"}normalized Laplacian {"} domain. We show that both luminance subtraction and amplitude division stages lead to significant reductions in redundancy relative to the original image pixels. We also show that the resulting quality metric provides a better account of human perceptual judgements than either MS-SSIM or a recently-published gain-control metric based on oriented filters.",
author = "Valero Laparra and Johannes Balle and Alexander Berardino and Eero Simoncelli",
year = "2016",
month = "1",
day = "1",
doi = "10.2352/ISSN.2470-1173.2016.16HVEI-103",
language = "English (US)",
series = "Human Vision and Electronic Imaging 2016, HVEI 2016",
publisher = "Society for Imaging Science and Technology",
pages = "43--48",
editor = "Pappas, {Thrasyvoulos N.} and {de Ridder}, Huib and Rogowitz, {Bernice E.}",
booktitle = "Human Vision and Electronic Imaging 2016, HVEI 2016",
address = "United States",

}

TY - GEN

T1 - Perceptual image quality assessment using a normalized Laplacian pyramid

AU - Laparra, Valero

AU - Balle, Johannes

AU - Berardino, Alexander

AU - Simoncelli, Eero

PY - 2016/1/1

Y1 - 2016/1/1

N2 - We present an image quality metric based on the transformations associated with the early visual system: local luminance subtraction and local gain control. Images are decomposed using a Laplacian pyramid, which subtracts a local estimate of the mean luminance at multiple scales. Each pyramid coefficient is then divided by a local estimate of amplitude (weighted sum of absolute values of neighbors), where the weights are optimized for prediction of amplitude using (undistorted) images from a separate database. We define the quality of a distorted image, relative to its undistorted original, as the root mean squared error in this "normalized Laplacian " domain. We show that both luminance subtraction and amplitude division stages lead to significant reductions in redundancy relative to the original image pixels. We also show that the resulting quality metric provides a better account of human perceptual judgements than either MS-SSIM or a recently-published gain-control metric based on oriented filters.

AB - We present an image quality metric based on the transformations associated with the early visual system: local luminance subtraction and local gain control. Images are decomposed using a Laplacian pyramid, which subtracts a local estimate of the mean luminance at multiple scales. Each pyramid coefficient is then divided by a local estimate of amplitude (weighted sum of absolute values of neighbors), where the weights are optimized for prediction of amplitude using (undistorted) images from a separate database. We define the quality of a distorted image, relative to its undistorted original, as the root mean squared error in this "normalized Laplacian " domain. We show that both luminance subtraction and amplitude division stages lead to significant reductions in redundancy relative to the original image pixels. We also show that the resulting quality metric provides a better account of human perceptual judgements than either MS-SSIM or a recently-published gain-control metric based on oriented filters.

UR - http://www.scopus.com/inward/record.url?scp=85011092689&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85011092689&partnerID=8YFLogxK

U2 - 10.2352/ISSN.2470-1173.2016.16HVEI-103

DO - 10.2352/ISSN.2470-1173.2016.16HVEI-103

M3 - Conference contribution

AN - SCOPUS:85011092689

T3 - Human Vision and Electronic Imaging 2016, HVEI 2016

SP - 43

EP - 48

BT - Human Vision and Electronic Imaging 2016, HVEI 2016

A2 - Pappas, Thrasyvoulos N.

A2 - de Ridder, Huib

A2 - Rogowitz, Bernice E.

PB - Society for Imaging Science and Technology

ER -