Posts: 2,051
Threads: 1,405
Joined: Jun 2011
[attachment=15470]
ABSTRACT
This paper proposes a new multimedia ontology based scheme for semantic multimedia data processing on the web. The ontology language ”Multimedia Web Ontology Language” (MOWL), is designed as an extension of OWL, the W3C recommended ontology language for the web. MOWL supports creation of and reasoning with perceptual modeling of concepts, and probabilistic evidential reasoning. Ontologies and context logic are closely related. Each context is ontology, and ontology inclusion could be one particular type of lifting axiom. An important aspect of context logic is that different contexts may be suitable for solving different problems.
Key terms -Multimedia systems, Ontology, Semantic Web, Bayesian network, Evidential reasoning
INTRODUCTION
The vision of semantic web proposes an environment where the data and services on the web can be semantically interpreted and processed by machines to facilitate human consumption. In today’s cyberspace, audio-visual artifacts compete with traditional text and data in their information content. Machine interpretation of multimedia data is therefore essential for realization of the semantic web vision. While textual documents are effectively processed by semantic web technology, application of the technology for multimedia data is still at its infancy. In this context, we propose a new ontology Based approach for contextual semantic interpretation of multimedia data.
Semantic web technology relies on ontology as a tool for modeling an abstract view of the real world and contextual semantic Analysis of documents. Ontology languages like Web Ontology Language (OWL)[1] uses linguistic constructs for modeling the real world and can be conveniently used for interpreting Textual documents. An attempt to use ontology for interpreting multimedia contents is hindered by the semantic gap that exists between media features appearing in the documents and the linguistic structures representing the concepts in the ontology. To cope up with this deficiency, there have been some attempts to extend ontology with addition of media examples for multimedia data processing [2]. However, such extensions do not meet the specific demands for reasoning with media data. Semantic retrieval in multimedia repository is generally carried out by intelligent concept recognition tasks that are specific to the media types and repository between the conceptual and the perceptual worlds is rather artificial. The key to semantic processing of media data lays in harmonizing the seemingly isolated conceptual and the perceptual worlds. Concepts are formed in human minds through a complex refinement process of personal experiences. Observations of the real world objects amounts to reception of large volumes of perceptual data through our sensory organs. The raw data go through a process of refinement to result in mental models.
PERCEPTUAL REASONING AND MOWL
Ontology is a formal description of the abstraction of a domain. Human beings use natural languages to communicate an abstract view of the world. Natural language constructs are symbolic representations of human experience and is close to the conceptual model that Semantic Web technologies deal with. Thus, it seems quite natural to use natural language constructs to represent the ontology elements. As a result, it becomes convenient to apply semantic web technologies in the domain of textual information. In contrast, media artifacts are perceptual recording of human experience. An attempt to use the conceptual model to interpret these perceptual records gets severely impaired by the semantic gap that exists between the perceptual media features and the conceptual world. However, the concepts have their roots in perceptual experience of human beings and the apparent disconnect between the conceptual and the perceptual worlds is rather artificial. The key to semantic processing of media data lays in harmonizing the seemingly isolated conceptual and the perceptual worlds. Concepts are formed in human minds through a complex refinement process of personal experiences. Observations of the real world objects amounts to reception of large volumes of perceptual data through our sensory organs. The raw data go through a process of refinement to result in mental models. The models are further abstracted over a large number of observations, to give rise to concepts, which are labeled with linguistic constructs to facilitate communication. The fact that concepts are abstractions of perceptual observations has an interesting consequence. A concept gives rise to the expectation of some perceptible media properties on its embodiment in a multimedia artifact. Observation of those media properties forms the basis of concept recognition in a multimedia document. Figure 1 depicts the formation of the concept Medieval Indian Monument and the abstracted visual patterns that are expected on an embodiment of the concept in a multimedia artifact. Note that the different instances of the concept have significant variations and the perceptual model comprises an abstraction of their common visual properties.
Fig. 1. Perceptual Model
Multimedia Web Ontology Language (MOWL) has been proposed to encode the domain knowledge pertaining to the expected perceptual properties of concepts and to reason with them, over and above the semantic properties that are encoded and reasoned with in OWL. Figure 2 depicts a small section of an ontology encodingmedia based description of concepts. The individual media patterns are often connected by some spatial or temporal relationships with each other, in context of a concept or an event. For example the dome of a medieval monument should occur above its other components, such as the minarets or the facade. moreover, the expected perceptual properties of concepts can be inherited by other concepts in
In Computer Science and information science ontology formally represents knowledge as a set of concepts within a domain, and the relationships between those concepts. It can be used to reason about the entities within that domain, and may be used to describe the domain.
In theory, ontology is a "formal, explicit specification of a shared conceptualization".[1] An ontology renders shared vocabulary and taxonomy, which models a domain – that is, the definition of objects and/or concepts, and their properties and relations.[2]
Ontology is a specification of a conceptualization. The word "ontology" seems to generate a lot of controversy in discussions about AI. It has a long history in philosophy, in which it refers to the subject of existence. It is also often confused with epistemology, which is about knowledge and knowing.
In the context of knowledge sharing, I use the term ontology to mean a specification of a conceptualization. That is, ontology is a description (like a formal specification of a program) of the concepts and relationships that can exist for an agent or a community of agents. This definition is consistent with the usage of ontology as set-of-concept-definitions, but more general. And it is certainly a different sense of the word than its use in philosophy.
What is important is what an ontology is for. My colleagues and I have been designing ontologies for the purpose of enabling knowledge sharing and reuse. In that context, ontology is a specification used for making ontological commitments. The formal definition of ontological commitment is given below. For pragmatic reasons, we choose to write ontology as a set of definitions of formal vocabulary. Although this isn't the only way to specify a conceptualization, it has some nice properties for knowledge sharing among AI software (e.g., semantics independent of reader and context). Practically, an ontological commitment is an agreement to use a vocabulary (i.e., ask queries and make assertions) in a way that is consistent (but not complete) with respect to the theory specified by ontology. We build agents that commit to ontologies. We design ontologies so we can share knowledge with and among these agents.
Multimedia Web Ontology Language (MOWL) has been proposed to encode the domain knowledge pertaining to the expected perceptual properties of concepts and to reason with them, over and above the semantic properties that are encoded and reasoned with in OWL. Figure 2 depicts a small section of an ontology encodingmedia based description of concepts. The individual media patterns are often connected by some spatial or temporal relationships with each other, in context of a concept or an event. For example the dome of a medieval monument should occur above its other components, such as the minarets or the facade. moreover, the expected perceptual properties of concepts can be inherited by other concepts in the domain, depending on their relationship. For example, a specific instance of a monument, the Taj Mahal, is made of a specific class of stones, marble, and therefore expected to ”inherit” its color and texture properties. This form of media property inheritance is quite distinct from classical property inheritance rules that is supported by classical ontology models and requires distinct form of reasoning. This is required for creating a complete ObservationModel of a concept using media properties of that concept and other related concepts.
