Visualization, artificial intelligence, digitization… These technological developments allow better understanding of works of art and are useful for their restoration, as well as for their authentication and their digital reproduction.
From the field archaeologist to the online art gallery, technology is a game-changer, especially in the field of art conservation, including study, restoration and conservation. In order to determine the nature of the materials used, the molding technique, the factors that threaten their integrity, the processings used before, even the existing drawings on which the artists “redrawn” and created their work, the curators resort to technologies that are always more advanced, using images and chemical analysis, whose data is processed by the software and, not surprisingly, used to train algorithms.
Artificial intelligence for better job identification
Imaging techniques (fluorescence, x-rays, infrared, spectral scanning, etc.) derived from those used by scientists are used to identify the materials of the work, the artist’s technique, “repaints” (front restorations), and any “underlayers”. As Sophie Lefebvre of C2RMF Communications (French Museum Research Center at the Louvre) explains: “These techniques allow you to dive deeper and deeper into the work and reveal elements invisible to the naked eye. Sometimes an artist would paint over a drawing or draw a shape different from the one he originally drew. Today we have more and more data, and more and more accurate data. » Then, the processing of these data allows you to install (in particular) the algorithms designed to study the received images.
Here’s how researchers at University College London used deep learning last September to reconstruct a Picasso painting (Crouching lonely nude), hidden under his famous canvas food for the blind. If X-rays made it possible to detect this “hidden” pattern and determine its composition, then AI allowed it to be reconstructed with the original colors. principle? Train the algorithm by “feeding” it dozens of Picassos from the same period (blue period) so that it “learns” the artist’s style and reproduces it. Once recreated, the work was 3D printed on canvas for a more realistic rendering.
In restoration, this approach, which combines radiography and artificial intelligence, offers new research opportunities that can help solve the challenges posed by artistic research. Testimony is Helene Dubois, project manager for the conservation of the Ghent Altarpiece, painted by the Hubert brothers and the Dutchman Jean Van Eyck. “This allows us to decipher complex technical images, diagnose superimposed or altered compositions, and damage that is not visible on the surface. We can better understand old master paintings, and the information disclosed helps experts protect and restore fragile works. »
Robotics and AI: European RePair project
Launched five months ago, the European project RePair (Reconstruction of the Past: Artificial Intelligence and Robotics) is helping archaeologists by combining artificial intelligence, visual recognition and robotics. “The AI-da robot is the first robotic artist with mechanical arms combined with a high-resolution scanner and AI-based 3D digital recognition software. He is able to instantly and autonomously collect fragments of objects of the past, broken into many parts. This method brings multiple offers that we confirm or not”, summarizes Marcello Pelillo, an artificial intelligence expert at Ca’ Foscari University in Venice and leader of the first project to use this new method: Pompeii, rich with numerous remains waiting to be preserved, restored or rebuilt. Thanks to the RePAIR project, two world-famous frescoes (the name has not yet been disclosed), made up of thousands of fragments, should be restored in the near future.
XpeCAM: A New Era of Spectral Imaging
“XpeCAM is an automated solution for restorers to collect data from a physical painting and understand what materials are present. For this, a spectral imaging device is used, which sends images to an artificial intelligence-based web platform.”, explains Antonio Cardoso, co-founder of Signinum, who developed this handheld spectral camera with XpectralTEK. At the Strasbourg-based Enterprise for Heritage Conservation (ECP), scientist Luke Ellerman uses the latest generation XpeCam 02, released in 2021. “The data obtained from the spectral analysis of the image is transferred to the software, which offers the restorer several options. The more scans and pictures, the more possible answers and the smarter the software becomes. Innovation is important because it is from data that we extract knowledge. » Currently, Luke Ellerman, who is working on a restoration of a contemporary work whose title he cannot reveal, admits that the original planned restoration protocol has been completely overhauled by data and analysis of these new technologies, resulting in much better options.
To get an accurate picture of the condition of the works and guarantee their authenticity, digitization is a pleasure for the art market players. “By releasing the ‘DNA’ of a work of art for sale, 5D digitization, by virtue of the accuracy of its data (support, repaint, etc.), gives the buyer even more assurance regarding the state of preservation. work and its authenticity. Data protection of a digitized work is carried out using a digital fingerprint, which makes it impossible to falsify the original », say Tajan, which was one of the first auction houses to use the technology developed by Swiss startup Artmyn. principle? The scanner photographs the work from all sides, and thousands of images are digitized and assembled according to the algorithm. This allows for an interactive image, an immersive video that can be manipulated in all directions and viewed under all types of lighting (direct, oblique, gliding, ultraviolet and infrared).
A godsend for art dealers who have been using this scanner in the Paris area more and more over the past year. “Wherever they are in the world, potential buyers can zoom in on the work, as well as move, rotate and manipulate it as if they were holding it in their hands – 5D images allow you to change perspective and tilt digitally. light emission. The quality is such that we are approaching the sensations of reality, including in terms of luminosity, texture and relief., says Gregoire Debuire, innovation manager at Artmyn. For Mathieu Fournier, Associate Director of Artcurial, “Artmyn and its innovative technologies invite us to the very heart of the paintings. This is a great opportunity to better understand and better experience the complexity of the masterpieces passing through our hands. This tool is a fantastic help in adapting to new applications in the art world.”.
From these high-resolution digitized images, digital content can be used for scientific purposes or for the general public, from virtual exhibition to augmented reality. And soon, as Sophie Lefebvre of C2RMF predicted, “Museums near work will definitely have a QR code that will provide the viewer with all the data that we now have in the laboratory. The viewer will be able to recognize the original and gain access to a wealth of knowledge that cannot be seen with the naked eye.”.