Learning no-reference quality metric by examples

Hanghang Tong, Mingjing Li, Hong Jiang Zhang, Changshui Zhang, Jingrui He, Wei Ying Ma

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

33 Citations (Scopus)

Abstract

In this paper, a novel learning based method is proposed for No-Reference image quality assessment. Instead of examining the exact prior knowledge for the given type of distortion and finding a suitable way to represent it, our method aims to directly get the quality metric by means of learning. At first, some training examples are prepared for both high-quality and low-quality classes; then a binary classifier is built on the training set; finally the quality metric of an un-labeled example is denoted by the extent to which it belongs to these two classes. Different schemes to acquire examples from a given image, to build the binary classifier and to model the quality metric are proposed and investigated. While most existing methods are tailored for some specific distortion type, the proposed method might provide a general solution for No-Reference image quality assessment. Experimental results on JPEG and JPEG2000 compressed images validate the effectiveness of the proposed method.

Original languageEnglish (US)
Title of host publicationProceedings of the 11th International Multimedia Modelling Conference, MMM 2005
Pages247-254
Number of pages8
DOIs
StatePublished - 2005
Externally publishedYes
Event11th International Multimedia Modelling Conference, MMM 2005 - Melbourne, VIC, Australia
Duration: Jan 12 2005Jan 14 2005

Other

Other11th International Multimedia Modelling Conference, MMM 2005
CountryAustralia
CityMelbourne, VIC
Period1/12/051/14/05

Fingerprint

Image quality
Classifiers
Metric
Image Quality Assessment
Classifier
Binary
JPEG2000
Prior Knowledge
General Solution
Learning
Experimental Results
Class
Training
Model

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Modeling and Simulation

Cite this

Tong, H., Li, M., Zhang, H. J., Zhang, C., He, J., & Ma, W. Y. (2005). Learning no-reference quality metric by examples. In Proceedings of the 11th International Multimedia Modelling Conference, MMM 2005 (pp. 247-254). [1385998] https://doi.org/10.1109/MMMC.2005.52

Learning no-reference quality metric by examples. / Tong, Hanghang; Li, Mingjing; Zhang, Hong Jiang; Zhang, Changshui; He, Jingrui; Ma, Wei Ying.

Proceedings of the 11th International Multimedia Modelling Conference, MMM 2005. 2005. p. 247-254 1385998.

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

Tong, H, Li, M, Zhang, HJ, Zhang, C, He, J & Ma, WY 2005, Learning no-reference quality metric by examples. in Proceedings of the 11th International Multimedia Modelling Conference, MMM 2005., 1385998, pp. 247-254, 11th International Multimedia Modelling Conference, MMM 2005, Melbourne, VIC, Australia, 1/12/05. https://doi.org/10.1109/MMMC.2005.52
Tong H, Li M, Zhang HJ, Zhang C, He J, Ma WY. Learning no-reference quality metric by examples. In Proceedings of the 11th International Multimedia Modelling Conference, MMM 2005. 2005. p. 247-254. 1385998 https://doi.org/10.1109/MMMC.2005.52
Tong, Hanghang ; Li, Mingjing ; Zhang, Hong Jiang ; Zhang, Changshui ; He, Jingrui ; Ma, Wei Ying. / Learning no-reference quality metric by examples. Proceedings of the 11th International Multimedia Modelling Conference, MMM 2005. 2005. pp. 247-254
@inproceedings{3e006131bd3d44108aa7ee191af5040e,
title = "Learning no-reference quality metric by examples",
abstract = "In this paper, a novel learning based method is proposed for No-Reference image quality assessment. Instead of examining the exact prior knowledge for the given type of distortion and finding a suitable way to represent it, our method aims to directly get the quality metric by means of learning. At first, some training examples are prepared for both high-quality and low-quality classes; then a binary classifier is built on the training set; finally the quality metric of an un-labeled example is denoted by the extent to which it belongs to these two classes. Different schemes to acquire examples from a given image, to build the binary classifier and to model the quality metric are proposed and investigated. While most existing methods are tailored for some specific distortion type, the proposed method might provide a general solution for No-Reference image quality assessment. Experimental results on JPEG and JPEG2000 compressed images validate the effectiveness of the proposed method.",
author = "Hanghang Tong and Mingjing Li and Zhang, {Hong Jiang} and Changshui Zhang and Jingrui He and Ma, {Wei Ying}",
year = "2005",
doi = "10.1109/MMMC.2005.52",
language = "English (US)",
isbn = "0769521649",
pages = "247--254",
booktitle = "Proceedings of the 11th International Multimedia Modelling Conference, MMM 2005",

}

TY - GEN

T1 - Learning no-reference quality metric by examples

AU - Tong, Hanghang

AU - Li, Mingjing

AU - Zhang, Hong Jiang

AU - Zhang, Changshui

AU - He, Jingrui

AU - Ma, Wei Ying

PY - 2005

Y1 - 2005

N2 - In this paper, a novel learning based method is proposed for No-Reference image quality assessment. Instead of examining the exact prior knowledge for the given type of distortion and finding a suitable way to represent it, our method aims to directly get the quality metric by means of learning. At first, some training examples are prepared for both high-quality and low-quality classes; then a binary classifier is built on the training set; finally the quality metric of an un-labeled example is denoted by the extent to which it belongs to these two classes. Different schemes to acquire examples from a given image, to build the binary classifier and to model the quality metric are proposed and investigated. While most existing methods are tailored for some specific distortion type, the proposed method might provide a general solution for No-Reference image quality assessment. Experimental results on JPEG and JPEG2000 compressed images validate the effectiveness of the proposed method.

AB - In this paper, a novel learning based method is proposed for No-Reference image quality assessment. Instead of examining the exact prior knowledge for the given type of distortion and finding a suitable way to represent it, our method aims to directly get the quality metric by means of learning. At first, some training examples are prepared for both high-quality and low-quality classes; then a binary classifier is built on the training set; finally the quality metric of an un-labeled example is denoted by the extent to which it belongs to these two classes. Different schemes to acquire examples from a given image, to build the binary classifier and to model the quality metric are proposed and investigated. While most existing methods are tailored for some specific distortion type, the proposed method might provide a general solution for No-Reference image quality assessment. Experimental results on JPEG and JPEG2000 compressed images validate the effectiveness of the proposed method.

UR - http://www.scopus.com/inward/record.url?scp=84891541368&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84891541368&partnerID=8YFLogxK

U2 - 10.1109/MMMC.2005.52

DO - 10.1109/MMMC.2005.52

M3 - Conference contribution

SN - 0769521649

SN - 9780769521640

SP - 247

EP - 254

BT - Proceedings of the 11th International Multimedia Modelling Conference, MMM 2005

ER -