Logarithmic profile mapping and Retinex edge preserving for restoration of low illumination images

Ngaiming Kwok, Haiyan Shi, Gu Fang, Ching-Feng Lin, Chin Yeow Wong, San Chi Liu, Shilong Liu, Md Arifur Rahman

Research output: Chapter in Book / Conference PaperConference Paperpeer-review

5 Citations (Scopus)

Abstract

![CDATA[Contents carried in an image are valuable information sources for many scientific and engineering applications. However, if the image is captured under low illumination conditions a large portion of the image appears dark and this heavily degrades the image quality. In order to solve this problem, a restoration algorithm is developed here that transforms the low input brightness to a higher value using a logarithmic mapping function. The mapping is further refined by a linear weighting with the input to reduce the un-necessary amplification at regions with high brightness. Moreover, fine details in the image are preserved by applying the Retinex principle to extract and then re-insert object edges. Results from experiments using low and normal illumination images have shown satisfactory performances with regard to the improvement in information contents and the mitigation of viewing artifacts.]]
Original languageEnglish
Title of host publicationProceedings of the 2016 9th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI 2016), 15-17 October, 2016, Datong, China
PublisherIEEE
Pages217-222
Number of pages6
ISBN (Print)9781509037100
DOIs
Publication statusPublished - 2016
EventInternational Congress on Image and Signal Processing -
Duration: 15 Oct 2016 → …

Conference

ConferenceInternational Congress on Image and Signal Processing
Period15/10/16 → …

Keywords

  • digital techniques
  • image processing
  • image restoration
  • logarithms
  • signal processing

Fingerprint

Dive into the research topics of 'Logarithmic profile mapping and Retinex edge preserving for restoration of low illumination images'. Together they form a unique fingerprint.

Cite this