When selecting pigments from a large set for restorative inpainting, it can often be challenging to create a mixture that will provide an exact match to the original artwork under a range of viewing and illumination conditions. In this research, a prototype computer program was developed that will aid the user by providing a color match and paint recipe that exhibits minimal metamerism when compared to the original artwork. The Gamblin Conservation Colors, a set of 43 colorants specially formulated for inpainting, were characterized in terms of their optical properties, absorption and scattering, according to Kubelka-Munk turbid media theory. Formulations were made using traditional spectrophotometric measurements and image-based measurements. The multispectral imaging system consisted of a trichromatic CFA camera coupled with two absorption filters; spectral reflectance data for each pixel location was estimated with a transformation based on calibration target images. Three targets were used for testing formulation accuracy: a target consisting of mixtures of Gamblin Conservation Colors, and two oil paintings. Pigment selection was reasonably successful, and good predictions resulted from both measurement techniques, but for more complex tasks such as pigment identification, a more rigorous colorant characterization approach may be needed. Predictions from image-based measurements were generally less accurate, and improvements in the camera model would likely remedy this. It is expected that this software will be of assistance to conservators by simplifying the process of selecting from a large set of available pigments, as well as reducing the possibility of damage to painted surfaces in cases where direct measurements are impractical. The open source nature of the software provides the opportunity for changes and addition of features in the future.

Library of Congress Subject Headings

Inpainting; Painting--Conservation and restoration; Image reconstruction; Color in art; Image processing--Digital techniques

Publication Date


Document Type


Department, Program, or Center

Chester F. Carlson Center for Imaging Science (COS)


Berns, Roy


Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: ND1650 .H34 2011


RIT – Main Campus