Images as Data

This talk tackles images as data in the digital humanities. This visual turn, enabled by a new set of computational methods such as image analysis using computer vision, foregrounds and disputes the logocentrism that has prevailed in digital humanities work over the decades. Abetted by a narrowly construed history of the field, textual computational work has often taken center stage, its longevity propelled by the centrality of particular disciplines to DH. We advocate for a broader role for the image, highlighting the potential of visual resources as the data for analysis and harnessing methods for analyzing images qua images. The talk is organized around three case studies: (1) Distant Viewing Protest Photography, (2) Photogrammar, and (3) Seeing Incunabula: Text as Image. They demonstrate how computational image analysis combined with institutional commitments to digitization along with inter and trans-disciplinary questions can forge avenues of inquiry in the field. We argue that thinking of images as data opens up a more capacious configuration of sources, evidence, and methods in DH.