Color correction for color distorted images

Frank Shih, Yan Yu Fu

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

1 Scopus citations

Abstract

Color distortion is less noticeable than noise and blur. Normally color distortion, especially found in the low bit-rate JPEG images, does not need to be corrected until it is above the level of JND (just noticeable distortion). The first part of this paper presents a comparative assessment of color distortion in different color spaces for blurred images. Experimental results show that color distortion increases as the magnitude of motion blur increases, and remains in the same level as the window size of Gaussian blur increases. The second part presents two color correction methods: One is RCC (Referenced Color Correction) and the other is NRCC (No-Reference Color Correction). Experimental results show that RCC can correct distortion by 88.3 % and NRCC can correct distortion by 60.71 %. The distorted images can be enhanced to have better perception.

Original languageEnglish (US)
Title of host publicationProceedings of the Third IASTED International Conference on Circuits, Signals, and Systems, CSS 2005
EditorsV.G. Oklobdzija
Pages206-211
Number of pages6
StatePublished - Dec 1 2005
EventThird IASTED International Conference on Circuits, Signals, and Systems, CSS 2005 - Marina del Rey, CA, United States
Duration: Oct 24 2005Oct 26 2005

Publication series

NameProceedings of the Third IASTED International Conference on Circuits, Signals, and Systems, CSS 2005

Other

OtherThird IASTED International Conference on Circuits, Signals, and Systems, CSS 2005
CountryUnited States
CityMarina del Rey, CA
Period10/24/0510/26/05

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Keywords

  • Color Correction
  • Color Distortion
  • Motion Blur

Fingerprint Dive into the research topics of 'Color correction for color distorted images'. Together they form a unique fingerprint.

Cite this