reSearch : enhancing information retrieval with images

Goodfellow, Martin Hugh and Hunt, Ela and McCafferty, Daniel (2009) reSearch : enhancing information retrieval with images. In: Technical Report, 1900-01-01. (Unpublished)

[img]
Preview
Text (strathprints008104)
strathprints008104.pdf
Accepted Author Manuscript

Download (503kB)| Preview

    Abstract

    Combining image and text search is an open research question. The main issues are what technologies to base this solution on, and what measures of relevance to employ. Our reSearch prototype mashes up papers indexed using information retrieval techniques (Terrier) with Google image search for faces and Google book search. The user can interactively employ query expansion with additional terms suggested by Terrier, and use those terms to expand both the text and image search. We test this solution with a selection of recent publications and queries concerning people engaged in research. We report on the effectiveness of this solution. It seems that the combination works to a large extent, as testified by our observations.