Vienna, 7 March 2017
CC-BY 4.0, Sandra Lehecka
University of Heidelberg, Germany
How would research in the humanities benefit if computers could handle images just as competently as they presently process text? What if performance would improve even further?
Presently, computers can process images only on a low level, such as pixel by pixel or based on textual annotations. Given vast amounts of visual data, there is, however, a dire need for a much more intuitive access to individual images and a more effective handling of large image corpora. After all, one of the main challenges we face is finding and setting related artifacts into perspective although they are in different, otherwise unrelated corpora. In short, the image analysis that is presently available to the humanities needs to be upgraded from the equivalent of a typewriter to that of a highly versatile content processor.
This talk discussed the challenges along the way towards artificial vision systems that can adequately support the humanities. Moreover, Professor Ommer analyzed the potential of machine vision for the humanities—and also the competences the humanities can offer to vision and machine learning.
ACDH Lectures are free and open to all. Registration closed.