Plenary lecture 11: semantic approaches for web multimedia processing and communication
Ashraf Ahmad · International Conference on Artificial Intelligence · 2009
The motivation for emerging the concept semantic web is due to the fact that search engines and text-based exploring are no longer adequate as these approaches involve an extensive information retrieval process. The deployed searching and retrieving descriptors are naturally subjective and their deployment is often restricted to the specific application domain for which the descriptors were configured. The new era of information technology imposes different kinds of requirements and challenges. Automatic extracted audiovisual features, as these features are more objective, domain-independent and more native to the audiovisual content are required. Therefore, this presentation is a detailed trail to introduce to the researchers, experts, students and practitioners very valuable references and guides to lea? them through their exploration and researching the multimedia content and semantic web. The results of leading researchers from around the world are successfully collected in one concise presentation in very semantic manner. This highly informative presentation includes three parts to construct a valuable reference resource for researchers, student and experts dealing with the emergence of integrated semantic multimedia field. The presentation introduces the concept of semantic web and multimedia content analysis to the readers through a logical sequence from standards and hypothesis through system examples, presenting relevant tools and methods. In addition, this presentation addresses approaches applied to specific problem domains such as content search, indexing and retrieval. The attendees may attain sufficient knowledge to start projects or researches related to the presentation theme from finding recent results and articles related to the active research area of integrating multimedia with semantic web technologies. The semantic web technologies are elaborated and ontology representation has been emphasized. The presentation provides an excellent summary of the fundamental theory behind applying a knowledge engineering approach to vision problems. This summery iconize the concept of semantic web and multimedia content analysis. A definition of the fuzzy knowledge representation that can be used for realizing on multimedia content applications has been provided with comprehensive analysis. The second part of presentation introduces the multimedia content analysis approaches and applications. In addition, some examples of methods applicable to multimedia content analysis are presented. Multimedia content analysis is very diverse and concerns many other research fields at the same time. Multimedia content analysis has strong diversity issues, as it starts from low level features such as (colors, DCT coefficients, motion vectors, etc.) up to the very high and semantic level (Object, Events, Tracks, etc.). The second part includes topics on structure identification (e.g. shot detection for video sequences), and object-based video indexing. These conventional analysis methods are supplemented by results on semantic multimedia analysis, including development and use of knowledge models for automatic multimedia analysis. The final part of the presentation discusses and describes example systems which current projects have been implemented, and include extensive results and real demonstrations. For example, real case scenarios like ECommerce medical applications and web services have been introduced. Natural language, speech and Image processing techniques and its application for multimedia indexing and contentbased retrieval have been elaborated with extensive examples and deployment methods.