Abstract

Viewing the change in reflected lights caused by changing illumination provides abundant information concerning illuminant and surface color properties.¹ A bilinear model describes, in simple viewing situations, how the quantum catches of a p-chromatic visual system are related to illuminant spectral power distributions and surface reflectance functions drawn from m- and n-dimensional linear models, respectively. Such models can be used to help recover illuminants and surface reflectances from quantum catch data when the visual system is presented v views of a set of s surfaces, where each view corresponds to a distinct illuminant. The choices of p,m,n,v, and s for which color structure can be recovered from chromatic motion follow from analysis of a set of conditions that a bilinear model must meet; single-view theory follows as a special case.

© 1992 Optical Society of America