Multispectral Analysis of Digital Images for Art Historical Research and Preventative Conservation of Paintings Research Project Concerning Non-Invasive Diagnostic Methods for the Preventative Conservation of Paintings.

2009 > 2014

Promotor(s) : Dr. Frederik Leen

Researcher(s) : Bruno Cornelis (Vrije Universiteit Brussel - ETRO), Prof. Dr. Ann Dooms (Vrije Universiteit Brussel - ETRO)

Research Team


  • Prof Dr Frederik Leen (Curator; Head of the Department of Modern Art)


  • Prof Dr Ann Dooms (Professor of Mathematics, Department of Electronics and Informatics (ETRO), Vrije Universiteit Brussel)


  • Bruno Cornelis (Engineer; PhD student, Vrije Universiteit Brussel)
  • Etienne van Vyve (Independent conservator; Instructor conservation and restauration, La Cambre, Brussels)
  • Pierre-Yves Desaive, Karine Lasaracina, Lies Van de Cappelle (Royal Museums of Fine Arts of Belgium, Digital Museum)


  • Prof Dr Ingrid Daubechies (Professor of Mathematics, Princeton University)
  • Prof Dr Peter Schelkens (ETRO, Vrije Universiteit Brussel)


  •  Lumière Technology, Paris (Multispectral imaging) 


Starting date: 01.01.2009
Ending date: indeterminate


The project generates an exploratory and experimental investigation in which new art historical and conservational research questions can be formulated pertaining to works of art, based on the analysis and processing of multispectral digital images of ultrahigh resolution and a wave-length area that reaches from infrared to visible light, to ultraviolet light. The results of very recent developments in digital data acquisition technology and image processing open up the opportunity to develop, on an experimental basis, a broad range of new possibilities and research questions for the art historical study and conservation of art works. Research material therefore involves images of a very high resolution with digital information that partially falls outside of the non-assisted optical spectrum. The interest of the project lies in the research into the possibilities hidden within the translation of content-based (i.e. art historical) research questions and material (i.e. conservational) issues into algorithms, and in the research into the relevance of new as yet unexplored information which such algorithms would bring to light. Dynamic interaction between the computer scientist, mathematician, the art historian and the curator/conservator is a primary concern for the project. The interest in terms of content lies in the design of new types of research questions which are made possible by this interaction and unique combination of multispectral data acquisition with state-of-the-art digital image analysis.