Examines the problems of developing a framework and testbed for quantitative assessment of image retrieval effectiveness. To better harness the extensive research on content-based retrieval and improve capabilities of image retrieval systems, this article advocates the establishment of common image retrieval testbeds consisting of standardized image collections, benchmark queries, relevance assessments, and quantitative evaluation methods. (AEF)