CHArt CHArt Computers and the History of Art - 1997 Conference Paper Abstract
This talk discusses the results of research conducted into the use of craquelure patterns on pictures as a form of identification and analysis. The techniques evolved for analyzing craquelure patterns have proved to be remarkably successful. Between 95 and 100% correct discrimination between pairs of art-historical categories were achieved for C14/15 Italian panel paintings, C15/16 Flemish panel paintings, C17 Dutch paintings and C18 French paintings. These results will be considered in relation to connoisseurial traditions reaching back to Mancini, Morelli and Wölfflin.
The representational method adopted in this research was derived from cognitive psychology and the analytical methods are those exploited in the social sciences. In parallel with this research an engineer has been attempting the mathematical representation and classification of craquelure patterns. His work should be completed by July 1997 and will be briefly discussed in this talk.
Spike Bucklow holds PG Dips in Artificial Intelligence and the Conservation of Easel Paintings. He has just been awarded a PhD (Cambridge) and is currently employed as the Research Scientist and Teacher of Theory at the Hamilton Kerr Institute. His current research interests are mainly medieval.