Faceted Metadata for Image Search and strongBrowsingstrong.pdfVIP

Faceted Metadata for Image Search and Browsing Ka-Ping Yee Kirsten Swearingen Kevin Li Marti Hearst ping@zesty.ca kirstens@ kevinli@ hearst@ Computer Science Division School of Information Management and Systems University of California, Berkeley University of California, Berkeley ABSTRACT want to browse and search images using textual category There are currently two dominant interface types for labels [1, 5, 7, 10]. Despite this, few image search engines searching and browsing large image collections: keyword- provide the ability to navigate images by rich category sets, based search, and searching by overall similarity to sample and those that do often have unwieldy interfaces [10]. images. We present an alternative based on enabling users We have developed an interface for large image collections to navigate along conceptual dimensions that describe the that allows users to navigate explicitly along conceptual images. The interface makes use of hierarchical faceted dimensions that describe the images [8]. The interface metadata and dynamically generated query previews. A uses hierarchical faceted metadata (described below) and usability study, in which 32 art history students explored dynamically generated query previews [14], seamlessly a collection of 35,000 fine arts images, compares this integrating category browsing with keyword se