Spectral Gamut Mapping Framework based on Human Color Vision.

摘要: A new spectral gamut mapping framework is presented. It adjusts the reproduction, choosing spectra within printer's that satisfy colorimetric criteria across a hierarchical set of illuminants. For most important illuminant traditional performed and for each additional considered colors are mapped into device pixel dependent metamer mismatch gamuts. computational separation method proposed in order to test framework. Utilizing this on seven channel printing system, experiments allowed deeper view structure gamuts possible directions color space which potential metameric transformation could map out-of-metameric-gamut colors. Introduction In recent years acquisition has become an active research eld. Today's technology able capture high resolution multichannel images with very small estimation error. This employed by museums artwork reproduction archiving applications. Wide-gamut multicolorant printers used management create accurate reproductions match originals under single illuminant. some applications it can be desireable multiple such cases needed. basic limitation physical ability devices reproduce re ectances. The printer much smaller than all natural lower bound dimensionality ectances determined through analysis databases [1]. Only looking onto difference does obvious majority cannot reproduced without error typical printer. becomes necessary unreproducible Such not unique optimal strongly depends special application. various metrics have been [2, 3]. To show advantage compared management, should one as visually correct other illuminants superior. An approach proposed, combining perceptual based corresponding three dimensional black [4, 5, 6, 7]. paper we presenting so matches original considering properties human vision. Spectral Gamut Mapping Framework Terminology explain use common terminology discrete spectra, resulting from sampling continuous at N equidistant positions visible wavelength range 380 nm 730 nm. Each ectance spectrum N-dimensional vector r ∈ [0,1]N adjusted vectors representing power distributions I1, . , RN following text observer's CIEXYZ tristimulus X(r, I) function = (r1, ,rN) illuminating I (I1, IN): 1 ∑i=1 ȳiIi ( ∑ i=1 xiIiri, ȳiIiri, ziIiri )T (1) where x, ȳ, z CIE matching functions 2◦ or 10◦ observer, respectively. nearly perceptually uniform CIELAB denoted L : 7→ inverse L−1 CIEXYZ. result value x called will M(x, {r | L(X(r, I)) x} (2) gamut, printable given device, G ⊂