Stauder, Jurgen, Sirot, Joel, Le Borgne, HervéORCID: 0000-0003-0520-8436, Cooke, Eddie and O'Connor, Noel E.ORCID: 0000-0002-4033-9135
(2004)
Relating visual and semantic image descriptors.
In: EWIMT 2004 - European Workshop on the Integration of Knowledge, Semantics and Digital Media Technology, 25-26 November 2004, London, UK.
This paper addresses the automatic analysis of visual content and extraction of metadata beyond pure visual descriptors. Two approaches are described: Automatic Image Annotation (AIA) and Confidence Clustering (CC). AIA attempts to automatically classify images based on two binary classifiers and is
designed for the consumer electronics domain. Contrastingly, the CC approach does not attempt to assign a unique label to images but rather to organise the database based on concepts.