Infoscience

Conference paper

Integrated browsing and searching of large image collections

Current image retrieval systems offer either an exploratory search method through browsing and navigation or a direct search method based on specific queries. Combining both of these methods in a uniform framework allows users to formulate queries more naturally, since they are already acquainted with the contents of the database and with the notion of matching the machine would use to return results. We propose a multi-mode and integrated image retrieval system that offers the user quick and effective previewing of the collection, intuitive and natural navigating to any parts of it, and query by example or composition for more specific and clearer retrieval goals

Related material