“Perceptually-Based Image Rendering of Art Paintings”

Digital archiving of art paintings was originally based on making simple color-calibrated images from the paintings [Martinez et al. 2002]. Practical digital archiving, however, is based on both shape information for describing surface geometries and spectral information for color reproduction. We showed that the surface shape of a painting was represented by a set of surface normal vectors of small facets that were estimated from camera data [Tominaga and Tanaka 2002]. Next, the spectral reflectance is more useful than color information, because an RGB image is device-dependent and valid for only the fixed conditions of illumination and viewing. Concerning the viewing conditions, most art paintings are hung on the wall indoors, which are often illuminated with incandescent lamps. Then we cannot neglect the effect of chromatic adaptation that is the most important color-appearance phenomenon of the human vision. In this sketch, we propose a method for perceptually-based image rendering of art paintings on a display device by considering surface properties and chromatic adaptation. Figure 1 shows the flow for the total digital archives. We describe the painting surface with surface normal, surface spectral reflectance, and reflection model. These surface properties are estimated using a multi-band imaging system. All the estimates are combined for image rendering.

Martinez, K., Cupitt, J., Saunders, D., and Pillay, R. 2002. Ten years of art imaging research. In Proceedings of the IEEE, vol. 90, 28–41.
Tominaga, S., and Tanaka, N. 2002. Measuring and rendering art paintings using an RGB camera. In Proceedings of EUROGRAPHICS, 299–306.