Catálogo de publicaciones - libros

Entertainment, education and training are changing because of multi-party interaction technology. In the past we have seen the introduction of embodied agents and robots that take the role of a museum guide, a news presenter, a teacher, a receptionist, or someone who is trying to sell you insurances, houses or tickets. In all these cases the embodied agent needs to explain and describe. In this paper we contribute the design of a 3D virtual presenter that uses different output channels to present and explain. Speech and animation (posture, pointing and involuntary movements) are among these channels. The behavior is scripted and synchronized with the display of a 2D presentation with associated text and regions that can be pointed at (sheets, drawings, and paintings). In this paper the emphasis is on the interaction between 3D presenter and the 2D presentation.

The growth of available entertainment information services, such as movies and CD listings, or travels and recreational activities, raises a need for personalization techniques for filtering and adapting contents to customer’s interest and needs. Personalization technologies rely on users data, represented as User Models (UMs). UMs built by specific services are usually not transferable due to commercial competition and models’ representation heterogeneity. This paper focuses on the second obstacle and discusses architecture for mediating UMs across different domains of entertainment. The mediation facilitates improving the accuracy of the UMs and upgrading the provided personalization.

The H. C. Andersen system revives a famous character and makes it carry out natural interactive conversation for edutainment. We compare results of the structured user interviews from two subsequent user tests of the system.

Public spaces are interactive by the very fact that they are designed to be looked at, walked around, and used by multitudes of people on a daily basis. Architects design such spaces to create physical scenarios for people to interact with, but this interaction will usually be one sided- the physical space does not usually change or react. In this paper we present three interaction design projects which add reactive dynamics into objects located in public spaces and in the process enhance the forms of interaction possible with them.

A simulation study about some basic dimensions of adaptivity that guided the development of the personalized summary reports about museum visits, as part of PEACH are presented. Each participant was exposed to three simulated tour reports that realized a sequential adaptive, a thematic adaptive and a non-adaptive version, respectively, and subsequently on each of the dimensions investigated. Results were unexpected. The possible reasons are discussed and conditions under which personalized report generators can be preferred over non personalized ones are proposed.

We describe a new prototype for a zoo information system. The system is based on RFID and allows to retrieve information about the zoo animals in a quick and easy way. RFID tags identifying the respective animals are placed near the animal habitats. Zoo visitors are equipped with PDAs containing RFID readers and WLAN cards. The PDAs may then read the RFID tag IDs and retrieve respective HTML-documents from a zoo Web server showing information about the animals at various levels of detail and languages. Additionally, the system contains a JXTA and XML based peer-to-peer subsystem, enabling zoos to share their content with other zoos in an easy way. This way, the effort for creating multimedia content can be reduced drastically.

We present a method for realistic, real-time simulation of human hair, which is suitable for the use in complex virtual reality applications. The core idea is to reduce the enormous amount of hair on a human head by combining neighbored hair into wisps and use our cantilever beam algorithm to simulate them. The final rendering of these wisps is done using special hardware accelerated shaders, which deliver high visual accuracy. Furthermore we present our first attempt for interactive hair styling.

The recognition of student’s motivational states and the adaptation of instructions to the student’s motivations are hot topics in the field of intelligent tutoring systems. In this paper, we describe a prototype of an Intelligent Chess Tutoring System based on a set of motivational strategies borrowed from Dweck’s theory. The main objectives of the prototype are to teach some chess tactics to middle-level players and to help them to avoid helpless reactions after their errors. The prototype was implemented using Flash Mx 2004. The graphical user interface encompasses a life-like character functioning as tutor.

We report on an effort to combine methods from storytelling and multimodal dialogue systems research to achieve flexible and immersive performances involving believable virtual characters conversing with human users. The trade-off between narrative control and character autonomy is exemplified in a game scenario.

A browsing system is described for transforming declarative web content into humorous-dialogue TV-program-like content that is presented through character agent animation and synthesized speech. We call this system Web2Talkshow which enables users to obtain web content in a manner similar to watching TV. Web content is transformed into humorous dialogues based on the keywords-set of the web page. By using Web2Talkshow, users will be able to watch and listen to desired web content in an easy, pleasant, and userfriendly way, like watching a comedy show.