Semantics-sensitive Retrieval for Digital Picture Libraries

Abstract

We present SIMPLIcity (Semantics-sensitive Integrated Matching for Picture LIbraries), animage database retrieval system, which uses high-level semantics classification andintegrated region matching based upon image segmentation. The SIMPLIcity systemrepresents an image by a set of regions, roughly corresponding to objects, which arecharacterized by color, texture, shape, and location. Based on segmented regions, the systemclassifies images into categories which are intended to distinguish semantically meaningfuldifferences. These high-level categories, such as textured-nontextured, indoor-outdoor,objectionable-benign, graph-photograph, enhance retrieval by narrowing down the searchingrange in a database and permitting semantically adaptive searching methods.