Date of Award

2013

Degree Type

Thesis

Degree Name

M.A.Sc.

Department

Electrical and Computer Engineering

First Advisor

Maev, Roman

Keywords

Communication and the arts, Applied sciences, Classification, Crack detector, Craquelure, K-nn, Lda, Morphology

Rights

CC BY-NC-ND 4.0

Abstract

Craquelure represents the unique crack formations that form on a material with age. This thesis is specifically concerned with the craquelure patterns found on historical art paintings. Although the significance of craquelure has been noted for over 300 years, recent research has shown that these patterns are related to the materials and methods employed by the artist and that these clues may assist in the task of attribution. It has been shown that different art paintings constructed from the same geographical location exhibit similar craquelure formation patterns. With the intention of alleviating expertise on the subject of craquelure, the development of a framework for the geographical analysis of craquelure patterns has been attempted in literature. This thesis seeks to expand on these results with the intention of increasing the accuracy rate in the classification of craquelure to their corresponding geographical origins. Through the use of mathematical morphology and various image processing techniques, craquelure images were converted to binary images. Specific features were then extracted from the binary image and used in the classification process. Several different classifiers were tested and compared in this thesis.

Share

COinS