Ontology-driven Annotation and Access of Educational Video Data in E-learning
Aijuan Dong, Honglin Li, Baoying Wang · InTech eBooks · 2010
The explosive growth of video data demands efficient and flexible access mechanisms. In this paper, we propose an ontology-driven framework for video annotation and video access. The goal is to integrate ontology into video systems in an effort to improve users' video access experience. The ontology-driven video annotation is a two-step process: video segmentation, and video annotation data extraction and organization. In video segmentation, we propose and utilize multi-mode segmentation procedures for presentation videos. In this procedure, the semantic-rich textual modality is integrated with the visual modality. To extract annotation data from videos and video segments, and organize them in a way that facilitates video access, we employ a multi-ontology based multimedia annotation model. In this model, a domain-independent multimedia ontology is integrated with multiple domain ontologies. The goal is to provide multiple, domain-specific views of the same multimedia content and thus meet different users' information needs. With extracted annotation data, we propose and implement ontology-driven video access. In ontology-driven video access, a user can select which ontology to interact with. The selection of ontology determines the set of annotation data and the group of relevant terms/concepts. As can be seen, ontology tailors the video access to users' domain-specific information access needs. To extend ontology-driven video to external heterogeneous data sources, web services are explored in this dissertation. Our experience shows that web service is an effective way to extract relevant documents from assorted, publicly available data sources. In this paper, we focus our discussion on presentation videos. But the general concept of ontology-driven video annotation and access is applicable to many other areas as well, for example, digital libraries, the Web and corporate video collections. To improve the work