TIB participates in interdisciplinary project on digital image search

German Research Foundation (DFG) grants € 500,000 in funding for joint project between Paderborn University and TIB

The German Research Foundation (DFG) has agreed to fund the project “Aby gets digital: ARAby – an adaptive retrieval and analysis tool to support image-oriented research processes”, initiated by Paderborn University and the Technische Informationsbibliothek (TIB) – German National Library of Science and Technology, within its e-Research Technologies Funding Programme. The externally funded project, with a total funding volume exceeding € 500,000, will run for three years. The aim of ARAby is to develop a computer-aided research tool for the better retrieval of digital images in humanities research.

The application for the cross-disciplinary joint research project, involving collaboration between computer scientists and cultural history experts, was submitted jointly by Professor Dr. Ralph Ewerth (TIB and Leibniz Universität Hannover) and Professor Dr. Eyke Hüllermeier and Professor Dr. Eva-Maria Seng, both from Paderborn University. ARAby seeks not only to exploit the methods offered by computer science and the techniques of digital information processing to explore phenomena in the cultural sciences, but also to generate completely new research areas, issues and processes from this cooperation.

The project was initiated by Professor Dr. Eva-Maria Seng, holder of the Chair of Cultural Heritage at Paderborn University. Since 2006, the Chair’s “Centre of Excellence for Cultural Heritage: Tangible – Intangible – Digital” has been operating a modern, digital image archive containing more than 10,000 images relating to the history of art and architecture. Owing to the ever-growing quantity of digital visual material, the idea arose to enable the automated, computer-aided search and analysis of large volumes of images.

An appropriate software tool will now be developed within the “ARAby” project to enable large volumes of image data to be used and assessed more easily in the humanities research process. This tool seeks to optimise image search in cultural heritage databases and to increase the performance capacity of existing scientific information systems relating to the research process in the humanities. Two computer scientists will contribute their expertise in machine learning, intelligent data analysis and image processing to the project: Professor Dr. Eyke Hüllermeier, holder of the Chair of Intelligent Systems at the Department of Computer Science at Paderborn University, and Professor Dr. Ralph Ewerth, Head of TIB’s Visual Analytics Research Group. “Our task will be to explore and develop algorithms for automatic image classification in this specific area,” stated Professor Dr. Ralph Ewerth, explaining the role played by TIB.

The starting point for the idea to develop such a tool was the method of systematic picture comparison developed by the art historian Abraham Moritz Warburg (1866-1929), known as Aby Warburg, at the beginning of the 20th century. Warburg used this method to compile his atlas of images entitled Mnemosyne. Back then, Warburg arranged all kinds of pictures on canvas-covered wooden panels to create a simultaneous overall view. By comparing this wealth of pictures and material through the ages, Warburg discovered recurring similarities and differences between formal and conceptual characteristics (known as pathos formulas). The montages he created not only served as his method of working, but also as his documentation of the results, simplifying the verification of hypotheses and leading to new questions being raised. The ARAby tool seeks to take Warburg’s method into the digital age and to expand it using the technologies offered by information science.