Catálogo de publicaciones - libros

In this paper we present a novel parameter learning and identification method of virtual garment. We innovate in the ordinary parameter identification process and introduce the fabric data (Kawabata Evaluation System data) to combine the expert’s knowledge with fuzzy system. With our method the parameters of virtual garment can be calculated automatically which are assigned by the animator’s experience and hard to tune in the past years. The statistic analysis and machine learning method are used to build the fuzzy system to present the fabric expert’s knowledge. With interactively inputting the human subjective variables to our method, the animator who knows little knowledge in physical attributes of fabric material can also create and tune the virtual garment application. The experimental results indicate that this method can be used in practical virtual environments and has the expansibility to other applications.

As shown in Paul Milgram et al’s Reality-Virtuality Continuum (1994), Augmented Reality occupies a very unique status in the spectrum of Mixed Reality. Unlike Virtual Reality, which is completely made up of the virtual and has been the most important theme for both research and practical applications, theoretical approach for Augmented Reality seems to be not keeping up with the technical development. In this paper, two qualities of Augmented Reality, which are virtuality of Augmented Reality and reality of Augmented Reality, are restated with reference to film and animation studies. In spite of the common and essential factor, computer generated images or virtual images, studies in computer vision appear to be set apart from other fields of studies. The object of this paper is to make an experimental conjunction for Augmented Reality and the concepts of Perceptual Reality, Second-order Reality, and possible world theory. It is expected to be useful particularly for studying artistic potentials and applications of Augmented Reality.

This paper presents a multimodal reference resolution approach in virtual environment, which is called RRVE. Based on the relationship between cognitive status and reference, RRVE divides the objects into four status hierarchies including pointing, in focus, activated, extinct, and step by step it processes multimodal reference resolution according to current status hierarchy. Also it defines a match function to compute the match probability of referring expression and potential referent, and describes the semantic signification and temporal constraints. Finally, sense shape is used to deal with the pointing ambiguity, which helps the user to interact precisely in immersive virtual environment.

This paper describes an ongoing research, SpaceCustomiser : InterActive, [SC : IA] which deals with development of digital design strategies based on non-Euclidean geometries, whereas the body in movement generates interactively architectural SPACE. The input – movement – is being electronically processed in such a way that the output represents a continuous, real-time modification of the space. For this purpose an on-site-built InterFace employing sensor/actuator technology enables translation of the recorded movement into spatial configurations. The InterAction between the body and the architectural space gives insight into, how the human body shapes space.

Designing and building Virtual Environments (VEs) has never been an easy task and it is therefore often performed by experts in the field. This problem arises even more when it comes to the design of behavior. To address this issue, an model-based approach called VR-WISE has been developed. This approach aims at making the design of VEs more accessible to users without or with little background in VE development. This means that the VE is specified by means of high-level models from which the actual VE can be generated. The purpose of this paper is twofold. Firstly, the design environment, supporting the approach, is described. We show how the models can be created in a graphical way. We also discuss the ”Verbalizer”. This module provides a natural language interface for the models, which enhances the intuitiveness of the specification. Secondly, we explain how behavior models can be specified.

In this paper, we propose a novel approach to modeling and dynamic creation of behavior-rich interactive 3D contents. The approach, called Beh-VR, enables dynamic generation of virtual scenes from arbitrarily selected sets of specifically designed reusable virtual objects, called VR-Beans. Behavior of the objects is determined by associated scripts encoded in a high-level language called VR-BML. The method significantly simplifies creation of interactive 3D scenes and can be applied to various application domains. Examples discussed in the paper are related to creation of educational and entertainment contents in the cultural heritage domain.

This paper describes the concept and the current state of development of "reactive spaces". How can one communicate with architectural spaces, through one’s movement or even one’s appearance? One goal of the project is it to develop a system which supports the interaction between spatial areas and its users. A space, which is changing with and through the user....which reacts to the user’s input. The basis for such experiments was laid in research fields such as Ambient Intelligence (AmI), Ubiquitous Computing or Hybrid Environments. If take these new communication and interaction possibilities some steps further, then it becomes clear that our traditional sense of space will be transformed radically by them.

This paper addresses issues of distributing multimedia presentations in an ambient intelligent environment, exams the existing technologies and proposes IPML, a markup language that extends SMIL for distributed settings. It uses a powerful metaphor of play, with which the timing and mapping issues in distributed presentations are easily covered in a natural way.

This paper presents the usability evaluation of our developed multi-sensory Virtual Assembly Environment (VAE). The evaluation is conducted by using its three attributes: 1) efficiency of use; 2) user’s satisfaction; and 3) reliability in use. These are addressed by using task completion times (TCTs), questionnaires, and human performance error rates (HPERs), respectively. A peg-in-a-hole and a Sener electronic box assembly tasks have been used as the task cases to perform the experiments, using sixteen participants. The experimental outcomes are analysed and discussed.

Soft shadow computation plays an important role in generating photo-realistic images. A new method of real-time all-frequency shadow rendering in 3D scenes under dynamic illumination is introduced in this paper. The non-linear wavelet approximation on the pre-computed radiance transfer matrix is used, which can help us to select the main directions sensitive to illumination of each sampling point. This approach efficiently avoids non-linear selection of the light vector in real-time rendering, and the approximation is used only once in the pre-computed phrase. We encode the resulting radiance transfer matrix into a sparse form, and store it as textures in 2D texture buffers. Meanwhile, the light vector is organized into a cube map on GPU to accelerate computation in the run-time process. This approach has been applied to render a variety of 3D scenes with soft shadows, and proved to be able to increase the relighting speed with no image quality reduction.