Improving Search in Social Media Images with External Information
The use of social media has increased considerably the recent years, and
users share a lot of their daily life in social media. Many of the users upload
images to photo-sharing applications, and categorize their images with
textual tags. Users do not always use the best tags to describe the images,
but add tags to get "likes" or use tags as a status update. For this reason,
searching on tags are unpredictable, and does not necessary return the result
the user expected.
This thesis studies the impact of expanding queries in image searches with
terms from knowledge bases, such as DBpedia. We study the methods
TF-IDF, Mutual Information and Chi-square to nd related candidates for
query expansion. The thesis reports on how we implemented and applied
these methods in a query expansion setting. Our experiments show that
Chi-square is the method that yields the best result with the best average
precision, and was slightly better than a search without query expansion.
TF-IDF gave the second best result with query expansion, and Mutual information
was the method that gave the worst average precision. Query
expansion with related terms is an exiting eld, and the information from
this thesis gives a good indication that this is a eld that should be more
explored in the future.
Place, publisher, year, edition, pages
Institutt for datateknikk og informasjonsvitenskap , 2014. , 147 p.
ntnudaim:9776, MIT informatikk, Informasjonsforvaltning
IdentifiersURN: urn:nbn:no:ntnu:diva-27329Local ID: ntnudaim:9776OAI: oai:DiVA.org:ntnu-27329DiVA: diva2:769303
Ramampiaro, Herindrasana, Førsteamanuensis