Browsing images using broad classification categories

Abstract

Browsing can serve an important role as a strategy for retrieving image information and can allow users to recognize what they are looking for more easily than describing it. This article describes a project which developed and tested an art history image data base which used broad classification units to group image sets into meaningful categories that support browsing. Categories from the Art and Architecture Thesaurus were used to provide broad classification units to guide users in browsing and searching. The system provided both browse and search capabilities and was tested with art history specialists and generalist users. The type of user and the use to which the image is put will influence the type of search strategy employed. Preferences for, and perfonnance in, the browse and search modes varied according to the user's familiarity with the subject domain. Both sets of users felt that the ability to use both browse and search in a multimodal system would best meet their search needs.