Intelligent Virtual World

The following sections are included:

• Contents

• Preface

For users of image management systems, and especially for the user who doesn't know what he wants until he sees it, these systems should be organized in such a way as to support intelligent browsing so that the user will be satisjed in the shortest amount of time. It is our belief that intelligent browsing should be mediated by the standard paradigms of image similarity as well as by an appropriate organization of metadata, including annotations and self-describing image regions.

Several applications in the areas of multimedia systems and computer graphics require to store and access large databases of images. Images are usually composed of 2D objects that contain many features (e.g., color, texture, shape, etc.) that have meaningful semantics. Among those, shape is an important feature that conforms to the way human beings interpret and interact with the real world objects. The shape representation of objects can therefore be used for their indexing, retrieval, and as a similarity measure. This chapter describes diferent image shape features that are used by the different shape representation techniques. Those features are used for organizing the useful shape information in index structures. In addition, we describe various methods for shape representation, and the multidimensional index structures that can be used to support shape retrieval queries. Finally, we describe several metrics and criterions that can be used to evaluate diferent shape representation techniques.

An ideal image retrieval system should retrieve images that satisfy the user's need, and should, therefore, measure image similarity in a manner consistent with human's perception. Unfortunately, perceptual consistency is very dfflcult to achieve, even for simple features such as color and texture. This paper summarizes current results of perceptual consistency and suggests possible future work in this direction. Striving for perceptual consistency should be a goal of the next-generation multimedia retrieval systems.

In this paper, we describe an image-based rendering technique called Inverse Concentric Mosaics (ICM), which is based on Concentric Mosaics (CM) (H.-Y. Shum and L.-W. He, 1999). While CMs allow scenes to be viewed from inside looking out, ICM permits the user to visualize objects from the outside looking in. ICM has the following advantages: viewpoints of objects can be changed in a continuous manner and at interactive speeds, photometric effects, if any, can be captured (enabling photorealism), and the image database can be easily acquired and processed. We show examples for synthetic and real objects.

This chapter presents several techniques for broadcasting multimedia data (audio and video) over the Internet. Internet broadcasting (also called webcasting) techniques have become very important in applications such as Internet (or Web) radio and television, real-time broadcasting of critical data (such as stock prices), distance learning, videoconferencing, and many others. We describe the current Internet broadcasting techniques including IP Unicast and IP Multicast, and we introduce a new technique IP Simulcast. The IP Simulcast approach is based on the hierarchical, binary structure of receivers, which at the same time become data senders or repeaters.

An overview on emerging IP traffic monitoring is presented. Important parameters to characterize the trafic, network, and QoS are discussed. The infrastructure and methodology to measure those parameters directly or to compute them based on other measurements are described. We also present a discourse on coping with the challenge of new transport architectures and technologies. In summary, a framework of IP traffic monitoring is presented.

This chapter describes our experiences with the design and implementation of a networked multimedia information management system in an object-oriented framework for distributed control applications, which has been achieved through a US Air Force project named HDIMI (Heterogeneous Distributed Information Management for the Infosphere) conducted by the University of Minnesota. The salient features of our framework are explained. Next, the chapter explores the challenges faced in integrating commercial off-the-shelf (COTS) object-oriented products into the framework, and how it affected the design and implementation. To demonstrate the capabilities of our system, we show a demonstration application for planning and monitoring air combat.

E-transactions via shopping agents constitute a promising opportunity in the e-markets. It becomes relevant what kind of information and what kind of exchange policies between competing sites can better model the e-market world. There are several steps for building e-business: (i) attracting the customer, (2) knowing how they buy, (iii) making transactions, (iv) perfecting orders, (v) giving effective customer service, (vi) offering customers recourse for problems such as breakage or returns, (vii) providing a rapid conclusion such as electronic payment. In the distributed e-market paradigm, these functions are abstracted via agents representing both contractual parts. In this chapter we will discuss the problem of contract negotiation in e-marketplaces. We succinctly present related topics commonly used to implement negotiation in e-markets, and analyze the bargaining process within e-markets. A study-case of a marketplace for e-services will be presented and analyzed.

On interconnected computer networks, we can communicate with each other using realtime media such as audio and video by distributed multimedia system that can integrate various realtime and discrete media data. The multimedia system is required to realize end-to-end quality of service (QoS) guarantee finctions. When users communicate with each other on interconnected different bandwidth networks, if the system can use translator or mixer functions that are defined by RTP, it will be able to guarantee more flexible QoS considering to wide band through narrow band networks. We have proposed Flexible Multimedia System (FMS) that is able to guarantee end-to-end QoS according to priority of parameters and consensus policy. In addition, the system is able to organize various multimedia services dynamically. The FMS is mobile agent based system, therefore the system will be able to organize translator or mixer dynamically. In this paper, we discuss about QoS guarantee functions using mobile agent and re-design our prototype system to use mobile transcoding functions for flexible multimedia communication.

In recent years, we have observed an increasing prevalence of small devices equipped with wireless network connection. Computing is no longer a discrete activity bound to the desktop. Current software targeted for these devices however is limited by the devices' computing power and small memory, which in turn restricts the functionality and complexity of the applications that the device can provide. As a result, the current software system is not flexible and adaptive enough to satisfy the high demand due to increased user population and sophistication in pervasive environments. To attack the problem, we propose a component-based software architecture that supports dynamic mobile code binding and loading. An application is made up of several small and mobile code components — called facets. Whenever a specific function is needed during the execution of an application, a facet that implements that function is selected and brought in from the network, executed and then discarded. As a result, small devices are able to run applications of any sophistication. Moreover, the selection of facets depends on the computing resources in the device, network conditions, and various environmental contexts. This makes our system more flexible and adaptive, and capable to enhance user mobility in a pervasive computing environment.

This chapter presents an approach to virtual training scenarios for developing virtual training systems. To overcome the difficulty in online instruction knowledge acquisition and verification, we use Petri nets theory as a tool for specifying and verifying varied training plans and corresponding training scenarios whereby the online instruction knowledge is derived. An experimental example, the Virtual Reality based Computerized Numeric Control (CNC) operations training system, shows that Petri nets based training tasks and training scenarios analysis can facilitate creating knowledge base of online training systems.

Learning via web is a trend of distance education. A typical web-based course presentation incorporating with multimedia technology is a HTML slide show with audio/video annotation. But, without detailed annotation on subjects of HTML slide discussed, learning efficacy is quite limited. Imposing additional guidance events upon HTML slide can improve the learning efficacy but incur complicated temporal, spatial, and content synchronization issues. In this chapter, a synchronized hypermedia lecture framework for web-based learning is presented and the related multimedia synchronization issues are addressed. A case study of implementation deployed in National Chi Nan University, Taiwan, for online English and Mandarin training is also presented.

This chapter discusses an educational computing platform, RIDEE (Real-time Interactive Distance Education Environment). RIDEE is designed based on high speed network, providing support for real time and interactive educational activities (distance lecture, distance seminar, etc.). RIDEE has been used in a regular distance lecture over JGN (Japan Gigabit Network) since 2002. In this chapter RIDEE's architecture, implementation, and experiments are described.

We propose a Question-answering (QA) system in Korean that uses a predictive answer indexer. The predictive answer indexer, first, extracts all answer candidates in a document in indexing time. Then, it gives scores to the adjacent content words that are closely related with each answer candidate. Next, it stores the weighted content words with each candidate into a database. Using this technique, along with a complementary analysis of questions, the proposed QA system can save response time because it is not necessary for the QA system to extract answer candidates with scores on retrieval time. If the QA system is combined with a traditional Information Retrieval system, it can improve the document retrieval precision for closed-class questions after minimum loss of retrieval time.

Because of its pervasiveness, the Internet is increasingly being used to support education and learning. We frequently observe a tendency to simply replicate traditional structures of classroom teaching and to make instruction accessible online. This approach adheres to what Barr and Tagg (1995) refer to as the Instruction Paradigm. We argue in favor of the Learning Paradigm and draw the reader's attention to critical considerations related to pedagogy and human learning that must guide the work of designing for learning. Effective learning must produce deep understanding, not merely knowledge reproduction. To this end, the Internet can be used to support learning by doing through the use of simulations, and online pedagogical agents can coach and scaffold students as they seek to master skills related to the domain of learning. In addition, students must engage in sense making and knowledge construction activities in environments that promote articulation and reflection. We use two learning environments to illustrate these ideas and to demonstrate more powerful ways of harnessing technology to support learning.

Recent advances on the Internet and Web technologies have enabled rapid development of web-based e-commerce systems. Since such systems are getting more and more complicated, AI techniques begin to play an important role for designing efficient and intelligent E-commerce systems. In this chapter, we, we give the overview of E-Commerce systems, and the significance of developing the intelligent web-based E-Commerce systems by using AI techologies: intelligent agent, CBR, game theory and fuzzy logic. We also present our work – an agent based negotiation model and applied CBR technique for data retrieval which are supported in the agreement and information phrases of E-Commerce transactions.

In the development of next generation e-commerce applications, it is necessary to provide a rich interaction for ensuring the higher customer satisfaction. This chapter concerns one of the important technologies which provide infrastructure for rich interaction, commonly known as personalization. We discuss how the personalization could be modeled and developed as a step to make e-commerce applications intelligent while enhancing the human-computer interaction in virtual environments. Our discussion will cover achieving personalization applications through Internet and also applications in augmented real world environments. In modelling personalization, two types of personalization called content personalization and system personalization are identified. The content personalization is achieved by using a virtual authoring mechanism described in detail in this chapter. To achieve expectation of system personalization, we use an intangible metaphor and a framework for application development, named as situation metaphor and Situated Information Filing and Filtering (SIFF), respectively.

The purpose of this chapter is to provide an introduction to and overview of the Quality of Service (QoS) protocols, which is now available or under development in order to provide services for the real-time multimedia applications in the Internet Protocol (IP) based networks. The goal of QoS support is to provide some level of predictability and control beyond the current IP “best-effort” service. Applications, network topology and policy dictate which type of QoS support is most appropriate for individual flows or aggregates. Accommodating the need for these diferent types of QoS support, a number of QoS protocols and architectures have been proposed for the past decade. Basically, the QoS architectures have evolved under two digerent routing control architectures: the traditional hop-by-hop IP routing control and the Multi Protocol Labeling Switching (MPLS) (International Data Corporation, 1999).

Both IntServ (Integrated Services) and DiffServ (Differentiated Services) are the new service architectures proposed to provide QoS on top of the hop-by-hop routing control of the current Internet. IntServ promotes end-to-end QoS guarantees by setting up an end-to-end connection for eachflow and maintaining the states of all connections, whereas DiffServ promotes scalability by pushing flow classification to network boundary and offering QoS to service classes. Targeting at both end-to-end QoS guarantees and scalability, a combination of the two architectures, an IntServ over DiffServ framework is also proposed. In contrast to the above approaches, MPLS routing is used to establish “fixed bandwidth pipes” analogous to ATM or Frame Relay virtual circuits. With MPLS, therefore, more of a “trafic engineering” approaches are deployed than QoS protocols.

In certain limited domains (e.g., distance education, e-commerce, e-conference), intelligent virtual societies already play an important role. The realization of the intelligent virtual world requires communication and intelligent technologies, which include multimedia information retrieval, multimedia communications, distance learning, e-commerce, and intelligent agents, In this era of information explosion that the Internet has brought about, it is imperative to exploit intelligent technology for efficient and effective search of information. Without significant adoption and advances in intelligent technology, the intelligent virtual world will remain a fantasy. Intelligent information search also requires accurate interpretations of the semantics of the acquired information. We provide an overview of research issues in several technology areas that enable realization of the intelligent virtual world.

The following sections are included:

• Author Index