Tools for Document Image Retrieval in Digital Libraries: the AIDI system

In the last few years, Digital Libraries became one important application area for Document Image Analysis and Recognition research. In this field, a relevant line of research is Document Image Retrieval (DIR) that aims at finding relevant documents relying on image features only. DIR techniques are used to index not only the textual content of a document, but also its layout, graphical objects, mathematical equations, and handwritten text. By integrating these capabilities with other traditional indexing approaches we expect it will be possible to define new search strategies that could be especially useful in scientific and technical collections. In this abstract we summarize our recent research on Document Image Retrieval techniques in the field of Digital Libraries that we integrated in the AIDI prototype system