A visual search engine for distributed image and video database retrieval applications

Abstract

This paper reports about an implementation of a search engine for visual information content, which has been developed in the context of the forthcoming MPEG-7 standard. The system supports similarity-based retrieval of visual (image and video) data along feature axes like color, texture, shape and geometry. The descriptors for these features have been developed in a way such that invariance against common transformations of visual material, e.g. filtering, contrast/color manipulation, resizing etc. is achieved, and that they are fitted to human perception properties. Furthermore, descriptors have been designed that allow a fast, hierarchical search procedure, where the inherent search mechanisms of database systems can be employed. This is important for client-server applications, where pre-selection should be performed at the database side. Database interfaces have been implemented in a platform-independent way based on SQL. The results show that efficient search and retrieval in distributed visual database systems is possible based on a normative feature description such as MPEG-7.