DIGITAL ANALYSIS OF PAPERS FOR THE AUTHENTICATION AND DATING OF ART

摘要

The digital analysis of the paper structure can play a majorrnrole in the authentication, dating and attribution of artrnwhich is very important for art historians. In this paper, arndigital analysis method is introduced for the extraction ofrnstructural features from machine-made papers. To generaterndigital images from these papers, we use a transmittedrnlight scanning method; since this method is feasible andrninexpensive as compared to other methods such as x-rayrnimaging. In machine-made papers, two types of structurernon the surface of a paper can be found: regular (periodic)rnand irregular (non-periodic). In this paper, we show thatrnthe power spectrum in Fourier domain is an adequate wayrnto extract structural features and isolate regular and irregularrnstructures. The structural features can be further usedrnfor authentication and dating of machine-made papers. Ourrnmethod is a simple and inexpensive method with respect torncosts and computational efficiency and it is invariant to therndirection in which the paper is scanned.