The explosion of digital data in the last two decades followed by the development of various types of data, including text, images, audio and video known as multimedia data. Multimedia Information Retrieval is required to search various type of media. There is comprehensive information need that can not be handled by the monolithic search engine like Google, Google Image, Youtube, or FindSounds. The shortcoming of search engine today related to their format or media is the dominance of text format, while the expected information could be an image, audio or video. Hence it is necessary to present multimedia format at the same time. This paper tries to design Unified Concept-based Multimedia Information Retrieval (UCpBMIR) technique to tackle those difficulties by using unified multimedia indexing. The indexing technique transforms the various of media with their features into text representation with the concept-based algorithm and put it into the concept detector. Learning model configures the concept detector to classify the multimedia object. The result of the concept detector process is placed in unified multimedia index database and waiting for the concept-based query to be matched into the Semantic Similarities with ontology. The ontology will provide the relationship between object representation of multimedia data. Due to indexing text, image, audio, and video respectively that naturally, they are heterogeneous, but conceptually they may have the relationship among them. From the preliminary result that multimedia document retrieved can be obtained through single query any format in order to retrieve all kind of multimedia format. Unified multimedia indexing technique with ontology will unify each format of multimedia.
|Number of pages||8|
|Journal||International Conference on Electrical Engineering, Computer Science and Informatics (EECSI)|
|Publication status||Published - Sep 2017|
- Concept-based search
- Information retrieval
- Multimedia information retrieval