Cinema and Machine Vision

This book explores the aesthetic, epistemic, and ideological dimensions of machine-seeing films and television using computers. It brings together scholarship in Film Studies and Artificial Intelligence by tracing their mutual interest in the automation of vision through technology, technically and historically. <italic>Cinema and Machine Vision</italic> critiques computational methods in the analysis of film and television, revisits key theories of the moving image and photography in the light of contemporary developments in AI and computer vision technologies, contributes to debates concerning early cinema, experimental cinema, and use of moving image archives, and theorises the emergence of synthetic media and generative approaches in computational humanities. The book presents this research programme in three layers: the first places computer vision in the wider context of technologies of vision by tracing its connection to photography and photographic practices; the second focuses on the computational analysis of time-based media and the computational production of new temporalities; the third makes the leap from using computer vision to analyse films to rethinking film criticism and scholarship computationally. Each layer is underpinned by practical computational experiments and a selection of aesthetics, film theory, and applied machine learning research. In aggregate, <italic>Cinema and Machine Vision</italic> opens for the reader key new problems that arise as AI becomes integral to visual culture, dispelling widely held assumptions about computer systems designed to watch and make images on our behalf and foregrounding how AI is profoundly reshaping the relation between the visible and the knowable.