Artificial intelligence and physics for art diagnostics: first results from “AIPAD” project

Book Heritage diagnostics is very demanding as the manuscript has the dual nature of material and textual object. The AIPAD project, a Next GenerationEU funded research, gives the proof-of-concept of novel methods for the diagnostics of ancient manuscripts bridging Artificial Intelligence (AI) and Physics. A noninvasive platform of image-based techniques and Deep Learning algorithms for acquiring the layered manuscript and for processing information is presented. Unconventional Thermal Quasi-Reflectograhy (TQR) in the mid-IR is integrated to multispectral imaging in the UV-VIS-IR for “delayering” manuscript features in surface-subsurface. As added value, Digital Image Correlation (DIC) allows to acquire structural information in full-field. In a cross-disciplinary approach with philologists, AI algorithms are used to process the image stacks that are annotated by the humanistic experts, e.g., to retrieve degraded text. First results from the Project are presented.