Inpainting, or virtual image completion, is sometimes the only way of restoring precious artwork. The ideas and tools of inpainting are described here for a broad readership. The real-world restoration of an illuminated manuscript shows the power of the technology and illustrates opportunities for technology development and art conservation.
The art of image restoration and completion has entered a new phase thanks to digital technology. Indeed, virtual restoration is sometimes the only feasible option available to us, and it has, under the name 'inpainting', grown, from methods developed in the mathematics and computer vision communities, to the creation of tools used routinely by conservators and historians working in the worlds of fine art and cinema. The aim of this book is to provide, for a broad audience, a thorough description of imaging inpainting techniques. The book has a two-layer structure. In one layer, there is a general and more conceptual description of inpainting; in the other, there are boxed descriptions of the essentials of the mathematical and computational details. The idea is that readers can easily skip those boxes without disrupting the narrative. Examples of how the tools can be used are drawn from the Fitzwilliam Museum, Cambridge collections.
Arvustused
'I've thoroughly enjoyed this outstanding book, which is destined to be the definitive source on inpainting but is also much more than that: it's a celebration of the creativity and beauty that are so often present in applied mathematics, and a reminder of the importance of human thought in art, science and their interplay.' Marcelo Bertalmío, Spanish National Research Council
Muu info
Explains, reviews and compares different inpainting imaging restoration methods in the challenging scenario of damaged medieval manuscripts.
1. Introduction;
2. Local inpainting methods;
3. Non-local inpainting methods;
4. Deep learning inpainting methods;
5. Methods inspired from cultural heritage;
6. Conclusions; Appendix. Manuscripts in focus; Bibliography; Index.
Simone Parisotto received his Ph.D. in Applied Mathematics in 2019, from the University of Cambridge. Following that, he served as Research Associate for the Fitzwilliam Museum in Cambridge. His research interests are in variational and numerical methods for inverse imaging problems, focusing on mathematical applications to heritage science. Patricia Vitoria received a Ph.D. in Computer Vision in 2021 from the Pompeu Fabra University in Barcelona. Her research interests include data-driven strategies for image processing and computer vision. She has been working on a wide range of image restoration problems such as inpainting, colourisation and deblurring. Coloma Ballester is a Professor of Applied Mathematics at Pompeu Fabra University in Barcelona. She works on theory and applications in image and video analysis, computer vision and applied mathematics. In the research field of image inpainting, she has made contributions in local methods, non-local approaches and deep learning strategies. Aurélie Bugeau is a Professor in Computer Science at the University of Bordeaux and at LaBRI. In 2022, she was appointed Junior Member of the Institut Universitaire de France. Her research interests include methods for image processing and analysis. Suzanne Reynolds, a Senior Curator at the Fitzwilliam Museum in Cambridge, has published widely on medieval and renaissance manuscripts and books. She directs the Fitzwilliam MINIARE project, Europa Nostra prize-winners for research (2023). Her most recent publication is the co-authored 'Pigments of Medieval British Illuminators: A Scientific and Cultural Study' (2023). Carola-Bibiane Schönlieb is a Professor of Applied Mathematics at the University of Cambridge. Her research is in mathematical image analysis and features a variety of interdisciplinary collaborations, among them with artists and art conservators on digital art restoration. She has been honoured by scientific prizes and invitations to give plenary lectures worldwide.
