Back to articles
Volume: 16 | Article ID: art00037_2
Four Spectral Channels for Natural Imaging of Scene Colors
  DOI :  10.2352/ISSN.2169-4451.2000.16.1.art00037_2  Published OnlineJanuary 2000

Ideal imaging channels should have the same spectral sensitivities of scene recording as vision channels and independently control the primary color sensations of eye. The trichromatic colorimetry is not helpful with its unreal primaries (XYZ) or alternating RGB color matching functions, which are little suggestive of non-negative channel sensitivities. Although the natural color space is formed by two alternating blue-yellow and green-red color axes, the nonnegative imaging space has to use an independent channel to render each of the four visual half-axes. The short-wave and long-wave channels in 56 modern photographic systems closely simulate their opponency against the yellow and bluish-green vision channels, which are roughly presented with a single middle-wave channel. The channel sensitivities are delimited by the neutral points of rod and cone. The necessary spectral overlaps of non-opponent channels are absent, thus decreasing the saturation of mixed colors. Positive dyes closely simulate the primary blue and red sensations, but the yellow and bluish-green ones are roughly replaced with an averaged yellowish-green color. An accurate rendering of natural saturation, metamery, the problematic bluish-green, purple-red, and low-illumination colors requires division of the middle-wave imaging channel into two separate ones. A spectral specification for the natural four-channel imaging is proposed, basing on the data of physiological, psychological, and medical studies of color vision.

Subject Areas :
Views 3
Downloads 0
 articleview.views 3
 articleview.downloads 0
  Cite this article 

Vitali V. Gavrik, "Four Spectral Channels for Natural Imaging of Scene Colorsin Proc. IS&T Int'l Conf. on Digital Printing Technologies (NIP16),  2000,  pp 560 - 565,

 Copy citation
  Copyright statement 
Copyright © Society for Imaging Science and Technology 2000
NIP & Digital Fabrication Conference
nip digi fabric conf
Society of Imaging Science and Technology
7003 Kilworth Lane, Springfield, VA 22151, USA