|Home \ Graduation Activities \ Post-Graduation Page||Login|
Análise e Identificação de Obras de Arte
Traditionally, the analysis of works of art was made only by historians and other experts, which is a long and complicated process, requiring the analysis of several evi- dences. Among these evidences are the analysis of the painter’s brushstroke, the analysis of painter’s signature and the analysis of the materials used. With the rise of tools for image processing and analysis, as well as signal processing, it was found that they could support the historians in the analysis of works of art, helping to solve many of the un- knowns of which still exist about some painters. Some museums have decided to create digital repositories from their collection so that researchers can work in these documents, thus avoiding physical contact with the work of art. Among those interested in this kind of tools are the museums and police departments. The museums want to authenticate and restore all their works of art. Police departments that fight smuggling and forgery of works of art are also interested because it would allow to validate the authenticity of confiscated works more quickly. This work considered two main phases, focusing on the works of painter Amadeo de Souza-Cardoso. In a first phase the signatures were analysed, by the application of image processing and pattern recognition techniques, in order to classify the authenticity and atribution using SIFT as well as to group similar signatures with the K-means method. The second phase was directed to the analysis of materials used in paintings starting from hyperspectral images of the paintings.
Start Date: 2011-09-20
End Date: 2012-11-02
Post-Graduation Student / Researcher / Professor: