Flamenco image browser: using metadata to improve image search during architectural design

Current user interfaces for image search are not successful at supporting architects' information access needs because they 1) do not recognize the context of architectural design, and 2) require complex textual queries that often result in either too many or too few images. This paper describes a new approach: using metadata about the images to aid search by generating dynamic query previews. Previews will help users visualize the contents of the collection, aid in making queries both more general and more specific, and help them to follow an information scent through a collection by giving them hints about where to go next. This paper describes a prototype system and plans for evaluation.