Digital image-based method to identify historical pigments in wall paintings

A smartphone image-based method for the colorimetric characterization of historical pigments in mural paintings

“…In this context, this paper describes the analytical method used by Sáez-Hernández and co-workers [ 6 ] to estimate the CIELAB values of a sample from the RGB values of an image captured with a colorimetrically characterized smartphone in order to distinguish among different historical inorganic pigments used in mural paintings using a chemometric approach.…”

“…The first step of the method requires to take an image of the sample under investigation. For the case reported in [ 6 ], the built-in camera application in the “pro” mode of Samsung Galaxy Edge S7 smartphone (model SM-G93F with a 12.2 MP camera sensor) was used. ISO was set to 100, aperture to 1/180 and white balance was 4000 K. The images were captured in .jpeg format.…”

“…In this article, all three SVC used the radial kernel (as it is needed to expand the classification boundaries further from two categories: in this case, three). This kernel uses an intrinsic value (ɣ) that also requires optimization, and for this case study, their optimal values were 0.5, 3 and 0.5 for yellow, blue and green, respectively (see the reference article for the discussion about the influence of each variable on predictive capacity [ 6 ]). To create these models, R [ 7 ] was used, and pROC [ 8 ], verification [ 9 ] and e1071 [ 10 ] packages were employed.…”