Font Size:
Ontology-based Collaborative Image Annotation
Last modified: 2011-12-18
Abstract
Interdisciplinary collaboration between computer scientists and archaeologists is enabling the development of innovative digital technologies in a variety of areas. Tracing Networks is a Leverhulme-funded research project that brings together archaeologists and computer scientists to investigate networks of crafts-people and craft traditions across and beyond the Mediterranean region, between the late Bronze Age and the late classical period. In recent years, archaeologists have gathered a massive amount of images, Cross-team knowledge sharing and analysis are vital for their research and being able to retrieve the right images, in the right context, and with the right level of confidence is essential.
We explore the use of semantic web technologies and introduce an ontology-based collaborative framework for image annotation, which allows users to tag concepts, relationships in the pictures and storing context regarding users. The framework also provides a systematic way to represent and combine uncertainty of statements as well as user-credibility measurement, which can be used for ranking search results. It addition, we are working on a GraphML-based query builder, which provides assistance to archaeologists who has difficulty with writing query for RDF data.
(short paper or poster)
We explore the use of semantic web technologies and introduce an ontology-based collaborative framework for image annotation, which allows users to tag concepts, relationships in the pictures and storing context regarding users. The framework also provides a systematic way to represent and combine uncertainty of statements as well as user-credibility measurement, which can be used for ranking search results. It addition, we are working on a GraphML-based query builder, which provides assistance to archaeologists who has difficulty with writing query for RDF data.
(short paper or poster)
Keywords
semantic web; ontology; tagging; uncertainty; credibility