Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Reflectances giving volumetric color constancy in daylight

Not Accessible

Your library or personal account may give you access

Abstract

The volumetric theory of color constancy1 provides illuminant-invariant estimation of reflectance spectra from reflected-light tristimulus values. The basis of the computation is the tristimulus volume ratio. Tristimulus volume ratios are illuminant-invariant when (a) the illuminant spectrum is a linear combination of N functions si(λ) free to vary only in their coefficients; and (b) each reflectance spectrum r(λ) is a linear combination of three basic functions rk(λ) plus a residual term R(λ) which is orthogonal to the 3N products fm(λ) of si and the tristimulus functions qj(i.e., r must be orthogonal to a forbidden subspace spanned by fm except for the projections of the functions rk). The color-constant reflectance spectrum r′(λ) closest (in a least-squares sense) to r(λ) can be obtained by subtracting off the forbidden-subspace components of r(λ). The present paper displays for N = 3 the r′(λ) for several real reflectances, based on si and rk drawn from principal-component analyses. We compute r′(λ) using the Gram-Schmidt orthogonalization on the forbidden-subspace functions. Each r′(λ) is compared with the function derived from the corresponding r(λ) by subtracting off the forbidden subspace of von Kries color constancy.2 The comparison favors the volumetric theory.

© 1985 Optical Society of America

PDF Article
More Like This
Heuristic analysis of von Kries color constancy

James A. Worthey and Michael H. Brill
TUJ8 OSA Annual Meeting (FIO) 1985

Color constancy and color discrimination

Brian A. Wandell
TUH3 OSA Annual Meeting (FIO) 1985

Surface reflectances and human color constancy

Jim M. Troost and Charles M. de Weert
WR8 OSA Annual Meeting (FIO) 1990

Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.