Painting by Numbers: Computational Methods and the History of Art

Leonardo Laurence Impett
2020
This thesis started its life as an EPFL fellowship in Computer Science, and ended -four and a half years later -in the newly-created School of Digital Humanities. This fluidity, and my own move away from my 'home' subject of machine learning, has made me entirely reliant on the advice, support, and (most of all) teaching of my colleagues and supervisors. This is probably the nature of cross-disciplinary, collaborative research: it rests entirely on having the right kind of collaborators. The
more » ... le project would have been unimaginable without its supervisors, Sabine Süsstrunk and Franco Moretti; both of whom have also been its coauthors. Sabine, not merely for her deep technical knowledge and advice, but for her imagination in cross-disciplinary visual computing (in conceiving, for instance, the angular pose model); and for having imagined the collaboration in the first place. And Franco, for having supplied much of the content (not least in operationalising); but also for patiently introducing an engineer to the research practices of cultural history. As collegial researchers, as supervisors, and as interdisciplinary scholars who are taken seriously by disciplinary communities, I could not have asked for more. Colleagues from across (and beyond) EPFL gave their friendship, advice, and invaluable intellectual input, in everything from neural network architecture to translating 15th century Russian. Amongst many others, I should at least name
doi:10.5075/epfl-thesis-7318 fatcat:d6eu5cwgwjfdxlbufrn7osabra