Digital analysis of papers for the authentication and dating of art

Abstract

The digital analysis of the paper structure can play a major role in the authentication, dating and attribution of art which is very important for art historians. In this paper, a digital analysis method is introduced for the extraction of structural features from machine-made papers. To generate digital images from these papers, we use a transmitted light scanning method; since this method is feasible and inexpensive as compared to other methods such as x-ray imaging. In machine-made papers, two types of structure on the surface of a paper can be found: regular (periodic) and irregular (non-periodic). In this paper, we show that the power spectrum in Fourier domain is an adequate way to extract structural features and isolate regular and irregular structures. The structural features can be further used for authentication and dating of machine-made papers. Our method is a simple and inexpensive method with respect to costs and computational efficiency and it is invariant to the direction in which the paper is scanned.