Rosa Ricci

The research aims at finding an optimized algorithm able to retrieve ornaments and emblems and extracts them from digitized databases. Historians usually employ thesaurus or subject-specific classification system (like Iconclass) to describe emblems and ornaments produced during the hand press period. Today the large amount of digitized items makes such manual description time consuming and expensive. Therefore, an automatic retrieval system is required. In the last years many solutions have been developed to optimize image retrieval and recognition in different research areas, but digital libraries still lack optimized solutions for their specifics needs. Using CBIR (Content Based Image Retrieval) it is possible to overcome different problems related to the comparison of different images as for example scaling invariance, complexity and scalability. The aim is to help historians and librarians to manage masses of data and to partially automatize image-based research.