Classification and Image Annotation for Bridging the Semantic Gap
Zurina Muda · ePrints Soton (University of Southampton) · 2007
The use of digital images is rapidly increasing in digital archives, community databases, as well as on the Web. This creates new challenges for image management and retrieval and promotes the importance of automatic image classification and annotation research. In general current content-based image retrieval methods are still struggling to deal with the semantic gap between low-level visual features and the high-level abstractions perceived by humans. Manual annotation is typically a difficult and tedious task involving a process of describing the content and context of an image to provide direct access to the semantics. Automatic classification can allocate images or image regions to specific object classes and automatic annotation also aims to add descriptive labels to images. This paper will explore classification and image annotation in bridging the semantic gap and present some related projects which illustrate the advantages of these techniques for image retrieval in the medical and cultural heritage domains.