AUTOMATIC HYPERMEDIA PRESENTATION GENERATION SYSTEM

Abstract

The widespread usage of multimedia data in recent years has led to an enormous interest in making these data available through the Web. Due to the complex requirements of multimedia data, the design and implementation of web-based information systems to fulfill the search, retrieval, and presentation necessities for such systems is not a trivial task. Design methodologies provide guidelines for the construction of multimedia information systems, so that the spatial, temporal and content-driven information retrieval requirements become manageable. Based on the separation-of-concerns principle some of these methodologies propose models to specify different aspects of the multimedia information system design. Initial work in this area carried out by introducing a standard reference model for the intelligent multimedia presentation systems (IMMPS). The standard reference model introduces a higher-level architecture and a plan-based approach to generate automated multimedia presentation. The model has led to realize the significance of knowledge representation and processing in producing adaptive and automated multimedia presentations. Multimedia information system design has been recently influenced by state-of –the-art technologies like the ones provided by semantic web. The semantic web technologies enable the system to effectively express and manipulate the knowledge over the Web. The knowledge representation and reasoning mechanisms of semantic web also facilitates generating personalised, adaptive and enhanced multimedia presentation through an automated process. This research introduces an information search and retrieval methodology that employs visions of semantic web for its data representation and reasoning tasks. Using a flexible and extensible data representation promotes application interoperability. The semantic relations and logical inferring is considered in the system for multimedia data processing and organisation in order to produce personalised and enhanced Web presentations. The research focuses on annotated multimedia data and multi levels of ontology in its design to address different aspects of an automatic hypermedia presentation generation based on the user’s query, devices, contexts and other preferences for targeted presentation. The system instead of providing a list of results, as what happens in the typical information search and retrieval systems, collects a list of candidate objects based on the relevancy to the query and also semantic relations iii between the objects and presents the results as a rich media-based presentation to the user. The ontologies are used to provide a shared understanding of the discourse domain, narration structure, and spatio-temporal concepts and entities. The main phases in the hypermedia presentation generation process could be categorised as follows: · the content collection phase, which searches the annotated objects addressed from different possibly heterogeneous sources. · the presentation structuring phase, which organises the collected data in a consequential scenario and also defines the composition of the hypermedia document. · the presentation generation phase, which constructs hypermedia presentations based on the previously integrated data. This dissertation concentrates on 1) defining a data model for multimedia data annotation, and 2) implementation of the different phases of an adaptive and automated presentation generation process in an integrated system, called MANA. The first part of the thesis focuses on section 1 (i.e. data annotation model), and literature review and design of the model. The second part is dedicated to literature review and design and architecture of an automated and semantic-enhanced hypermedia presentation generation system. Chapter 1 discusses the research questions and provides an outline of the dissertation content. Chapter 2 gives an overview of a multimedia data annotation and description structure. The chapter reviews some of the existing models for multimedia data annotation and representation. Chapter 3 describes our proposed data model based on the requirements described in chapter 2. It discusses that the suggested model which is represented based on standard vocabularies and document representation structures, is an adequate model to represent annotated multimedia data. The annotation structure also considers the role of domain and contextual knowledge, and reasoning in processing the different aspects of multimedia data. The chapter concludes that the proposed model is suitable for representing multimedia data for the purpose of a semantic supported hypermedia presentation generation process. Chapter 4 analyses the semi-automatic and automatic presentation methodologies embodied in the existing systems. Chapter 5 describes the authoring requirements for an automated hypermedia presentation generation system. The chapter concentrates in particular on the data collection, narration structuring, and presentation generation phases to generate an automatic hypermedia presentation in response to the user’s query and preferences. Chapter 6 describes the MANA system implementation iv and shows how the system satisfies the stated requirements. Chapter 7 provides a conclusion and discusses the future work. The implementation of the MANA system demonstrates how semantic support can be applied to our multimedia data annotation model. We show that the model can be represented satisfactorily through RDF/XML representations. We describe how implicit data hidden in the relations between the objects are used to extract explicit knowledge about the objects and organise the objects in a presentation scenario in response to information queries. The domain ontology and the narration ontology are employed to specify high-level knowledge about the discourse domain and the presentation structure. We emphasize the contribution of these knowledge representations to the automatic presentation generation process. The research describes the integration of existing models of multimedia annotation and representation into a richer document model and discusses implementation of an automated hypermedia presentation system, and also demonstrating the feasibility of the implementation.