A Taxonomy of the Image: On the Classification of Content for Image Retrieval

Abstract

Image database (IDB) systems are at present often designed to test technology and the efficacy of retrieval algorithms, rather than being oriented towards delivering functionality to users. Research is necessary to design interfaces geared towards human usage of images. The starting point of this research needs to be consideration at a fundamental, user-centred level of how people perceive and interpret images. This article considers literature from many disciplines to describe a taxonomy of image content, from direct sensory elements to high-level abstractions. The nine categories derived will later be validated and used to direct the design of visual query interfaces for IDB systems.