Published July 5, 2011 | Version v1
Video/Audio Open

Babel revisited: a taxonomy for ordinary images indexing in a bilingual retrieval context

  • 1. McGill University, Montreal


With the large volume of digital images now accessible on the World Wide Web, users
searching for images can be overwhelmed by many factors. Too many available images,
images indexed with an incomprehensible vocabulary or one that is too specialized to be
useful are but a few examples of issues leading to frustration. In addition, language barriers
still prevent Web users from retrieving the images they need.
This contribution presents the preliminary results of a study proposing to explore the
behaviours of image searchers from four different linguistic communities. The purpose of
this preliminary study is to examine queries formulated by image searchers to learn about
the terminology used and evaluate how this terminology can be eventually incorporated
into the development of a bilingual taxonomy for digital image indexing. Forty participants
from four different linguistic communities (English, French, Chinese and Russian native
speakers) were asked to write the queries they would use to retrieve ten images that were
shown to them consecutively. Then they were invited to fill out a questionnaire on their
behaviours as an image searcher on the Web.
The results of this research allowed the acquisition of knowledge of user terminology
standards and an assessment of how that terminology might be integrated in the develop-
ment of a bilingual taxonomy for improved indexing of ordinary digital images. Moreover,
since language barriers regularly prevent users from easily accessing information of all kinds, the
bilingual taxonomy will constitute a clear benefit for image searchers who are not overly
familiar with images indexed in English, which is still the dominant language of the Web.



Files (9.4 MB)

Additional details