The agglomeration of data in Arachne over the last three years led to
the insight that the manual labour of creating metadata for every image
seemed impractical. Between 2002 and 2007 the number of images stored
in Arachne expanded from 50,000 to 200,000; our present stock of 820,000
images represents an increase of around 400%. In cooperation with the
German Archaeological Institute (DAI) in Rome, the “image_grid project”
is seeking to solve these problems.
With
the “image_grid project” we have automatically contextualised about
400,000 images from the photo library of the DAI Rome. Since its
founding in 1928, the photo library of the DAI Rome has become one of
the world ́s most im- portant archives for archaeological imagery and
many scientists from all over the world visit it, in order to study
these images.
The images were digitized using old microfche-slides and the only
information was a printed index. We have created a digital image
database using OCR, in order to take the structure of this index and by
using the negative numbers of the images. We have also linked images
automatically with complex metadata, which are derived from digitized
inventories, scientifc reference works, images, which are now already
stored in Arachne and other sources. Negative numbers are also used to
connect images with OCR-texts (as in the iDAI.Bookbrowser or
GoogleBooks), and by extracting knowledge from these books via
text-mining we can connect new metadata to these digital images.
0 Comments