SEMEDIA Showcase - Research Applications
Search Environments for Media
PhotoSearch (Yahoo!)
Overview
PhotoSearch is an application for searching photos. It is a keyword-based search application that searches Flickr photos based on metadata information added by the photo owners. The main features of PhotoSearch are various tools for assisting the user in the search process. The tools allow the user to disambiguate ambiguous queries, expand queries with related terms, and to get metadata-similar images for an example image.
Features
The PhotoSearch application has the following main features:
- • Query disambiguation: Given an ambiguous query -- e.g., Washington -- PhotoSearch will suggest disambiguation options -- e.g., D.C. or Seattle. See furthter Figure 1.
- • Query refinement: Given a query -- e.g., Barcelona -- PhotoSearch will suggest query refinement terms for either making the query more general -- e.g. Spain or Catalonia -- or more specific -- e.g., Sagrada Familia or Barcelona architecture. See further the TagExplorer showcase and Figures 2 and 3.
- • In-context exploration of search results: For the query refinement/disambiguation terms, PhotoSearch helps the user to visualise the effect of the refinement by highlighting the search results that match the refinement criteria. See further Figure 3.
- • Query by example: Given a query and a photo: PhotoSearch will suggest a query for retrieving similar photos. See further Figures 1, 2, and 3.
The search functionality of PhotoSearch is powered by the Flickr API.
Suppose the user decides to browse photos from London (Figure). The user is of course shown the top photos relevant to her query (Figure-E). Additionally, the TagExplorer shows a cloud of related tags, such as England, United Kingdom, Southwark, City, Big Ben, London Eye, Thames, party and travel (Figure-C). The tags England, United Kingdom, Southwark, and City are grouped together as locations; the tags Big Ben, London Eye, and Thames are grouped together as subjects; and the tags party and travel are grouped together as activities (Figure-C).
Given the tag-cloud of related tags the user can continue her browsing by choosing to add to her query a term from the tag cloud (click on the respective green plus) or start a new query based on one of the tags in the cloud (click on the tag). Given the new query a new tag-cloud is generated and the photo list is updated. For example, suppose the user chooses to add the tag Thames to her London query. The tag will now show different tag-cloud relevant to the new query, including the tags Tower Bridge, bridge, river, South Bank, etc. Now the user can again choose to add tags to her query or browse other tags. This process can then -- in principle -- continue forever.
Demonstration
Click image for real-time demo
Links
