Unprecedented sources and methodologies make it possible to work on such questions. Images can be extracted from an unprecedented and massive corpus of digitized sources, and they can be described and studied with Deep Learning applied on pattern recognition, text mining, historical geomapping and the methodologies of global history, art history, visual studies, and cognitive sciences. For the period since the 1900s to the arrival of the internet, digitized news and photo magazines, catalogues and journals related to art, artworks and posters, and digital images can be crossed with press reviews and archival material, to explain how artistic and media images circulated over Europe and how they have helped to disseminate styles and visual patterns, as well as new ideas about politics, the body, or nature. It is thus possible to identify the most recurrent images in Europe of the past century, their cognitive features, the characteristics of their authors; to study their impact in copies, imitations, visual quotations, their channels of circulation and their contribution to European cultural homogenization. While doing so, we will question the old geopolitical model of prescriptive centres and imitative peripheries; and we will offer an epidemiology of images -an innovative model for the analysis of iconographic diffusions that can be applied across periods, languages, and disciplines, and later on, for subsequent projects, on visual circulation on the Internet.