How are semantics used in V4Design?

What is the role of Semantic Web?

According to World Wide Web Consortium (W3C) “Semantic Web is an extension of the World Wide Web which is associated with the vision to make published data easily-accessible from machines”. In semantics, data are represented under subject–predicate–object format, which is known as triples. Subjects are resources. Predicates or properties are the means to connect a subject to an object and show the relationship between them. Objects may be resources, values or empty nodes. Some popular semantic languages that have been developed are Resource Description Framework (RDF) and Web Ontology Language (OWL). Ontologies, on the other hand, offer the discipline and structures under which the data can be mapped semantically. Categories, properties and relations between the concepts, data and entities are described in each ontology which are associated with a specific domain of knowledge. Semantically represented data are saved into RDF triplestores, named Knowledge Bases.

What about Linked Data?

A common definition of Linked Data is the following: “Linked Data is structured data which is interlinked with other data so it becomes more useful through semantic queries”. The main reason behind that is that since data are represented semantically, serious problems, like heterogeneity of data, are overcome. DBpedia is a popular technology that offers a wide amount of entities. Each entity is described by various properties. DBpedia data is shared via the web into an RDF format and can be easily retrieved via semantic queries to further enrich already existing information. So, the answer to the question “Should I publish a dataset on the web?” is definitely yes!

How are semantics used in V4Design?

The results of V4Design project are strongly associated with the role of semantics, as they combine the information coming from heterogeneous sources into asset structures. Each video (Basic Asset) is associated with video-related, text, shot detection and aesthetics information. Text analysis results are associated with DBpedia entities and enriched with DBpedia information. The basic asset information contain for instance the location, latitude, longitude and textual description. Each 3D model (Full Asset) is associated with the results of 3D model reconstruction, Style Transfer, Scene Recognition, Spatio-Temporal Building & Object Localization, Text Generation and Basic Asset. The full asset information contains for example thumbnail URL, date of reconstruction for the 3D model, style of the stylized model etc. Such information is provided in V4Design platform and the V4Design Virtual Reality tool. In such way, we help architects and game designers to have all the information needed for each 3D model, without the need to search for this information on the web.