Modeling the adapted illumination color considering the degree of chromatic adaptation

Chromatic adaptation transforms (CAT), originally proposed by von Kries and subsequently refined over a decade, have become increasingly comprehensive. However, these models were derived from corresponding color data obtained in object color matching tasks, not investigating the adaptation to illumination color, which directly determines the color tone of the illuminated scenes. This study conducted a series of cross-media color matching experiments specifically designed to evaluate chromatic adaptation to ambient illumination across a wide range of illuminants varying in chromaticity and luminance. Observers matched the white point of the reproduced image on the display under dark viewing conditions to the perceived tone of real-world scenes. The results revealed that existing CAT models, which primarily employ luminance-based adaptation factors, fail to predict the adaptation state of the ambient illumination accurately under these conditions. To address this limitation, we developed a revised CAT that incorporates both the chromaticity and luminance of the illumination condition. The findings not only advance our understanding of adaptation mechanisms but also offer practical implications for perceptually consistent white-point adjustment in auto-white-balance (AWB) algorithms for digital imaging.