Automated Canvas Analysis for Painting Conservation


By Brendan Tobin 1. Motivation Distinctive variations in the spacings between threads in a painting's canvas can be used to show that two sections of canvas originated on the same bolt of fabric. A canvas spacing match between two works by the same artist can provide strong evidence toward the authenticity of a painting. Figure 1. shows a matching canvas spacing pattern between a set of three paintings by Nicolas Poussin. Analysis of the structure of a painting’s canvas can also identify areas of a painting susceptible to damage, or that may have been modified in the past. The individual threads in a canvas are visible in an x­ray radiograph of a painting, however additional processing is necessary to visualize and compare canvas structure data. Current canvas analysis techniques as described in Erdmann et. al. are not capable of resolving thread spacings at the level of individual threads. This work seeks to develop thread­level canvas analysis techniques to identifying the location of all thread crossings in a x­ray radiograph image using machine learning techniques to train a classifier capable of identifying thread crossings in a radiograph. The position of individual thread crossings can then be used to develop a thread level­map of canvas structure.

1 Figure or Table

Cite this paper

@inproceedings{TobinAutomatedCA, title={Automated Canvas Analysis for Painting Conservation}, author={Brendan Tobin} }