MaxRGB Reconsidered

$25.00 plus tax (Refund Policy)

Buy Article:

Abstract:

The poor performance of the MaxRGB illumination-estimation method is often used in the literature as a foil when promoting some new illumination-estimation method. However, MaxRGB has usually been tested on images of only 8-bits per channel, where clipping of high radiances is likely to have occurred. The question arises as to whether the method itself is inadequate, or rather whether it has simply been tested on data of inadequate dynamic range or with inadequate preprocessing. In particular, is MaxRGB's underlying assumption that there is a white or white-equivalent surface present in every scene too strong? This question is explored here in two ways. The first avenue of investigation is based on a new database of 105 sets of multiple-exposure images. High-dynamic range images are constructed from these sets as well. The color of the scene illumination is determined by taking an extra image of the scene containing four Gretag Macbeth mini-Colorcheckers placed at an angle to one another. MaxRGB is found to perform surprisingly well when tested on either the multiple-exposure or the high-dynamic range images. The second avenue of investigation is to add some simple preprocessing to the basic MaxRGB algorithm. By removing clipped pixels followed by median filtering, MaxRGB also performs better than previously reported when tested on test images of common color constancy test sets, specifically the Simon Fraser University 321-image indoor set. In particular, the Wilcoxon signed-rank test indicates that MaxRGB outperforms the most recent bright-pixel variant of color by correlation on the 321 set. MaxRGB is also competitive against the recent Edge-Based algorithm and significantly better than the computationally intensive Bayesian method on the Grayball set and the Colorchecker set. Overall, the results presented demonstrate that MaxRGB is far more effective than it has been reputed to be.

Document Type: Research Article

DOI: http://dx.doi.org/10.2352/J.ImagingSci.Technol.2012.56.2.020501

Affiliations: School of Computing Science, Simon Fraser University, 8888 University Drive, Burnaby, British Columbia V5A 1S6, Canada

Publication date: March 1, 2012

More about this publication?
  • The Journal of Imaging Science and Technology (JIST) is dedicated to the advancement of imaging science knowledge, the practical applications of such knowledge, and how imaging science relates to other fields of study. The pages of this journal are open to reports of new theoretical or experimental results, and to comprehensive reviews. Only original manuscripts that have not been previously published, nor currently submitted for publication elsewhere, should be submitted.
  • Editorial Board
  • Information for Authors
  • Subscribe to this Title
  • Membership Information
  • Information for Advertisers
  • Terms & Conditions
  • Privacy Policy
  • ingentaconnect is not responsible for the content or availability of external websites
Related content

Tools

Favourites

Share Content

Access Key

Free Content
Free content
New Content
New content
Open Access Content
Open access content
Subscribed Content
Subscribed content
Free Trial Content
Free trial content
Cookie Policy
X
Cookie Policy
ingentaconnect website makes use of cookies so as to keep track of data that you have filled in. I am Happy with this Find out more