We are creating more images than ever before. Photographsand their lookalikes saturate our field of vision with information and stimuli. The sheer volume of images readily available online opens up new possibilities for image processing through large-scale computation—a fusion of sensor data visualizations, photography, computer simulations, and AI hallucinations. These photographic doppelgängers are not merely snapshots captured by a camera; they are technical artifacts produced by larger-than-human-scale computational processes. I call them Hyperimages.
Hyperimage theory views the recent computational turn in photography as a continuation of photographic history and narratesit through the concept of scale. Scale as the contemporary milieu in the age of overabundance of images, big data, and large vision models. Scale, as in large-scale deep learning models that extract statistical patterns from the mass of photography readily available online. Scale as an act of imaging, revealing aspects of the world by resolving a frame from the space-time continuum, or sampling patterns from training data. Behind the apparent visual abundance, however, scalability appears to displace the humanistic pursuit of meaning, truth, and beauty in image-making. Examining hyperimage practices through the lens of scale, the thesis articulates the aesthetic politics behind the datafication of photographic seeing.