The Semantic Gap in Image Retrieval

When?

Wednesday 7 February 2007, 14:00 to 15:00

Where?

39BB02

Open to:

Staff, Students

Abstract:

The semantic gap is often referred to in papers on image retrieval or multimedia information handling. However, whilst many authors have been happy to make reference to it, few have attempted to characterize it in any detail. The talk will describe an AHRC project which has attempted to do this, in particular in the context of image retrieval. It will summarise current attempts to begin to bridge the gap both through developments in content-based techniques, the application of semantic web and knowledge technologies and recent progress in auto image annotation.

The talk will briefly present research into the way picture searchers articulate real queries and how they are typically resolved through a combination of traditional metadata and the knowledge of the searcher. The second part will review progress in content-based image retrieval and automatic annotation in recent years and explore the types of query that they are able to address. The semantic gap between features that can be extracted directly from images and the semantics that the human searcher attaches to the visual information will be revisited and various staging posts across the gap will be identified. such as raw data to features, features to objects, objects to labels and labels to semantics. The ways in which these sub gaps can be bridged in some instances will be discussed and our research on hard and soft auto annotation described.

 

Notes:

Paul Lewis is a professor in the Intelligence, Agents, Multimedia Group within the School of Electronics and Computer Science at the University of Southampton. His main research interests are currently concerned with the broad area of multimedia knowledge management, in particular addressing problems in image and video processing and analysis, multimedia annotation and semantic description of media. He is particularly involved with designing and developing novel facilities for multimedia information retrieval, navigation and browsing with applications in both the medical and the cultural heritage domains. His research is building on ideas from low-level media processing, knowledge management and emerging semantic web technologies.