In this paper an attempt has been made to develop a decision tree classification based algorithm for craquelure identification in old paintings. Craquelure can be an important element in judging authenticity, artist's workshop as well as for monitoring the environmental influence on the condition of the painting. Systematic observation of craquelure will help to build a better platform for conservators to identify cause of damage, thus a proper tool for precise detection of the pattern is needed. However, the complex nature of the craquelure is a reason why an automatic detection algorithm is not always possible to implement. The result presented in this work is an extension of known semi-automatic technique based on a region growing algorithm. The novel approach is to apply a decision tree based pixel segmentation method to indicate the start points of craquelure pattern. This, in particular applications may improve significantly the overall effectiveness of the algorithm. © 2013 Springer-Verlag.
CITATION STYLE
Gancarczyk, J. (2013). Decision tree based approach to craquelure identification in old paintings. In Advances in Intelligent Systems and Computing (Vol. 184 AISC, pp. 11–17). Springer Verlag. https://doi.org/10.1007/978-3-642-32384-3_2
Mendeley helps you to discover research relevant for your work.