A study of multilingual social tagging of art images

Abstract

The goal of this study is to compare social tagging patterns in two languages in image collections of art, while seeking exploitable strengths for the application of multilingual social tagging in digital libraries and museums. Crowdsourcing the annotation of digital image collections of artworks to different language communities has the potential to bridge language borders and reach wider audiences. This mixed methods study is based on a collection of digital images of paintings for which tags in Spanish and English were collected. The results show that the level of agreement in the vocabulary describing an image does not change significantly when adding a second language, but different cultural perspectives can be found for certain images when comparing less frequent tags across languages. Understanding and comparing tagging behaviors across languages is necessary for the design of user interfaces that support diversity and encourage sharing of perspectives about the artwork images.