Macro-level similarity measurement in VizIR

This paper analyzes the similarity measurement in Content-based Image and Video Retrieval systems (CBIR). The goal is to identify preliminaries for successful queries as the basis for the implementation of a query engine in the Content-based Visual Information Retrieval framework (VizIR). VizIR is an open CBIR framework for researchers, software developers and instructors. Past efforts in CBIR have lead to several generalpurpose prototypes. However, these prototypes differ in implemented feature classes, user-interfaces and similarity measurement. VizIR aims at overcoming this unsatisfactory situation. The paper overviews wide-spread techniques for similarity measurement in CBIR, derives a general querying model and proposes conditions for similarity measurement algorithms on the macro-level. Based on these conditions two methods (the Linear Weighted Merging method and the Query Model approach) are evaluated and the superior method chosen for the VizIR project. Additionally, the major goals of the VizIR project are outlined and interested researchers are invited to participate in the project.