The Descriptive Camera is a prototype camera designed and created by Matt Richardson, a photographer and computer programmer who teaches a Computational Camera and Photography class at New York University. The camera does not produce actual photographs. Instead, it prints out a text description of the item “photographed.” This is achieved via the transmission of an actual photographic representation to Amazon’s Mechanical Turk service. It is an actual human at Amazon’s outsourced center who views the image, composes a brief text description of the image, and transmit the text back to the camera’s assigned printer.
Nick Bilton, writing for the New York Times, took a picture using the Descriptive Camera of the outside of building in New York City. The printout he received from the camera stated: “This is a faded picture of a dilapidated building. It seems to be run down and in the need of repairs.”
Mr. Richardson views the project as primarily experimental, but notes several practical applications as well. The Descriptive Camera could be used by visually impaired persons, as well as for entertainment and categorization or archiving purposes.
The device is a prototype, and Richardson estimates that the camera’s constituent parts cost roughly $200.00. Each use of the Mechanical Turk service costs $1.25.
Computational cameras and photography is an emerging field of both academic study and practical application. The Massachusetts Institute of Technology offers a class similar to the NYU class and describes a computational camera as attempting to “digitally capture the essence of visual information by exploiting the synergistic combination of task-specific optics, illumination, sensors and processing.”