AM Posters 2009 START Conference Manager    

User-assigned preferred entry point levels for Web image searching compared with Pyramid model

Hyuk-Jin Lee and Diane Neal

(Submission #86)


Abstract

Few studies have been conducted to identify the image attributes needed by users for various photographs genres. Each object has one particular level at which contact is made first with semantic memory. The notion of entry point is an attribute of individual exemplars rather than an attribute of categories. This study investigates the differences on recognizing the entry points of image according to different photo genres in the Web environment. We created the hierarchy having 7 main categories and total 28 subcategories by analyzing the most popular Flickr tags to analyze the results. We discuss if our categories regarding entry points can be classified or extended into the 10 levels of the image attributes of Pyramid structure model, and investigate whether this model can serve as an entry point model in providing access to Web images. Findings from this research contributes what are the preferred entry points of various genres of photographs used for Web searching, and the outcome may contribute on enhancing Web image (photograph) information retrieval performance including the possibility of application of Web 2.0 techniques.

File(s)

[Additional materials, author bios (DOC)]  

  
START Conference Manager (V2.54.6)