0%
As interactive media services evolve with the internet, there is a growing demand for applications where users can actively control their perspectives and experiences. To support such interactivity, not only audiovisual content but also contextual information such as “when,” “where,” and “how” the content was recorded becomes crucial. In response, we have designed the Software Defined Media (SDM) Ontology to comprehensively describe spatio-temporal media and enable its structured use.
The SDM Ontology provides a flexible vocabulary for describing video, audio, and sensor data—along with recording equipment, processing systems, and associated semantic metadata. During classical and jazz concert recordings, we collected multi-angle videos and multi-channel audio, annotated them using the SDM Ontology, and published them as Linked Open Data (LOD). This approach allows not only performers or creators but also third-party developers to reuse and reinterpret the data across various applications.
A prominent application built on this ontology is “Web360²,” a browser-based interactive viewer. It dynamically retrieves 360-degree videos and spatial audio through SPARQL queries to the published LOD and lets users experience events from arbitrary viewpoints and listening positions. User studies confirmed that the application offers intuitive operation and immersive engagement, making it promising for education, arts, and entertainment.
This research contributes a foundational structure for enabling richly contextualized and extensible media experiences. Looking forward, we envision SDM Ontology playing a key role in AI-assisted content understanding, urban data archiving, and bridging physical and virtual environments in the era of the metaverse.
digital twins extended reality
digital twins
autonomous driving machine learning
machine learning v2x
autonomous driving v2x
extended reality
v2x
machine learning v2x
