Representing and Perceiving Depth in Digital Images

Back to Funding History

Elodie Fourquet, Assistant Professor of Computer Science, and her collaborator Flip Phillips, Professor of Psychology & Neuroscience at Skidmore College, have received a two-year award of $101,262 from the Picker Interdisciplinary Science Institute for their project “Representing and Perceiving Depth in Digital Images”. Humans easily and reliably perceive relative depth in two-dimensional images, such as photographs, realistic paintings, and even sketches. We do this despite the inaccurate depiction or absence of many depth cues that exist in natural 3D environments. Can we understand and digitally reproduce the visual information that makes this possible?

To produce 2D images of 3D scenes, this project uses computer graphics algorithms based on the projective geometric calculations of Renaissance artist-mathematicians. Psychophysical methods measure observers’ perception of the 3D scenes as the 2D pictorial information varies, providing critical insight into the human visual system’s interpretation of 3D space. The results of these experiments will guide improvement of the depiction of 3D depth in 2D imagery. This, in turn, will facilitate richer content in computer-generated images for illustration and visualization — two important research areas in computer graphics.