No-Reference Image Quality Assessment Based on Image Naturalness and Semantics

Runze Hu*, Wuzhen Shi, Yutao Liu, Xiu Li

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

Automatically providing feedback about the quality of natural images could be of great interest for image-driven applications. Toward this goal, this paper proposes a novel no-reference image quality metric capable of effectively evaluating the image quality without requring the information of the original image. The proposed method delivers a comprehensive analysis of the image quality through exploring its statistical natural properties and high-level semantics. Specifically, we adopt the NSS regularities based method to characterize the image naturalness, and take the semantic information of the image through the deep neural networks. The quality prediction model is then derived by SVR to analyze the relationship between these extracted features and the image quality. Experiments conducted on LIVEC and CID2013 databases manifest the effectiveness of the proposed metric as compared to existing representative image quality assessment techniques.

Original languageEnglish
Title of host publicationDigital TV and Wireless Multimedia Communications - 18th International Forum, IFTC 2021, Revised Selected Papers
EditorsGuangtao Zhai, Jun Zhou, Hua Yang, Ping An, Xiaokang Yang
PublisherSpringer Science and Business Media Deutschland GmbH
Pages203-214
Number of pages12
ISBN (Print)9789811922657
DOIs
Publication statusPublished - 2022
Externally publishedYes
Event18th International Forum of Digital Multimedia Communication, IFTC 2021 - Shanghai, China
Duration: 3 Dec 20214 Dec 2021

Publication series

NameCommunications in Computer and Information Science
Volume1560 CCIS
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference18th International Forum of Digital Multimedia Communication, IFTC 2021
Country/TerritoryChina
CityShanghai
Period3/12/214/12/21

Keywords

  • Deep neural network (DNN)
  • Image quality assessment (IQA)
  • Naturalness

Fingerprint

Dive into the research topics of 'No-Reference Image Quality Assessment Based on Image Naturalness and Semantics'. Together they form a unique fingerprint.

Cite this