Project Overview

Exploration and Visualization of the disjunction between curator-generated descriptive tags and those created by computer vision algorithms. Elaboration of where computer vision falls short in describing digital representations of art objects due to assumptions inherent to the vision system and what the computer wants to see in terms of texture, field of view, likelihood. Analysis of the presentist and popular ways of seeing embedded in machinic and distant viewing practices.

Data Sources and Methods

Results: Individual Image Attribute Tags and Captions Compared

Direct Link: Fusion Table

Chart View Highlighting Computer Vision Uncertainty

See also: Fusion Table

Results: Summary Tags Info

See also: Direct Link

Results: Summary Tags: Curator Tags

See also: Direct Link

Results: Summary Tags: Machine Tags

See also: Direct Link

Discussion