Comparing user‐generated and librarian‐generated metadata on digital images
Marija Petek · OCLC Systems & Services · 2012
Purpose Images can be seen in a different way by different users. The purpose of this paper is to examine how users describe images and to ascertain whether differences exist between users and librarians in creating metadata on images. Design/methodology/approach The paper compares metadata on digital images generated by users to metadata generated by librarians. A sample of images taken from Digital Library of Slovenia and Flickr is presented to students to assign tags. The tags are grouped into categories and classes of attributes and compared to keywords added by Slovene librarians and to tags created by Flickr visitors. Findings The number of assigned tags differs greatly among survey participants, librarians and Flickr users, the participants being the most productive. A majority of tags reflect perceptual attributes and tagging is mostly done for personal benefit. The matching rate for all images is 41.4 percent; matching is a little higher with the Flickr images. Practical implications Social tagging can be used to develop control vocabularies reflecting users' language and to provide access to digital images. Originality/value The paper presents quantitative data on image attributes used by users in describing images.