This paper presents an autostereoscopic 3-D workbench which can be enjoyed in home PC environment. Autostereoscopic 3-D workbench is a system where the viewer can interact with virtual 3-D objects directly with his or her hands. To achieve realityenhanced direct manipulation 3-D image is presented with the 3-D visual display FLOATS, which enables presentation of realistic 3-D objects within the reach of the viewer. Since the conventional system requires large cost and space, however, it cannot be set up in home PC environment. In the present paper the author proposes a compact optical system which requires neither large space nor much cost to realize autostereoscopic 3-D workbench.

抄録全体を表示

We have suggested an entertainment application system for children named as "Tangible Playroom". It has been designed as a basic system for future computer entertainment system. It uses a large floor screen, string based force display, "SPIDAR" and real-time physics simulator to realize our concepts, usual space, direct manipulation and emotional world. We have also developed a content software for Tangible Playroom named as "Penguin Hockey". It is a simple hockey game playing with A.I. penguins and snowman pucks. It has been presented for children's event. As the result, about 400 children had played it. This paper reports its development and technical details and results of questionnaire from players.

抄録全体を表示

In order to bring out a potential ability of a broad range of visual representation on computing systems for games and entertainment, user-friendly interfaces and a wide range of variations in novel interaction techniques are greatly expected. This paper proposes a novel system called "Kirifuki" which has an ability of operating a computer by breathing in and blowing out. The principle of this system is as follows; the computer screen is projected onto the physical desk, and when the user blows onto it, the visual objects on the screen can be manipulated by breathing without using hands and any of physical pointing devices. In this paper, we describe our prototype system and demonstrate its application sub-systems which especially applied to entertainment fields. We also have conducted a preliminary evaluation of our system and the result shows that our system can operate in the usual indoor environment.

抄録全体を表示

In this paper we present a method to synthesize virtual human motion in realtime computer graphics. In order to make a motion sequence, we prepare basic short motion segments and concatenate them with transformation for satisfying both internal and external constraints. We introduce a fast algorithm for motion transformation utilizing Spacetime Constraints with small amount of computation by restriction of external constraints and representation of motion segments with characteristic points.

抄録全体を表示

Media performance unit cell/66b use monochrome images and minimal sound to expand the performance potential of the body, leading to a more essential bodily expression. Here, the relationship of the performing bodies with the visual imagery and soundscape, becomes an ever-shifting dynamic. Currently, cell/66b and the author are developing wearable devices to synchronize body gestures to the visual and sound network. When devices are used in the performing arts, the way in which body movements are assigned to images and sound is an important matter. Apparent relationship does not necessarily lead to expanding the potential of the performance. In weak bindings, for example where a viewer perceives the relation of cause and effect but cannot notice the mapping, the viewer is induced to invoke his or her own images.

抄録全体を表示

Video motion capture is useful for computer animation because we do not need expensive sensing systems to capture complex motion. However the conventional motion analysis techniques can be failed due to the partial occlusions. In this paper, we propose a robust tracking method by minimizing both intensity change and joint torque applying to the body parts along the long image sequences. We also developed the object-oriented software systems working on many platforms.

抄録全体を表示

A horseback archery (called Yabusame in Japanese) is the traditional art of Japan that dates from the Heian Era (10th century). Since a horseback archery requires both horse riding skill and archery skill simultaneously, ordinary people cannot experience this excitement during their lifetimes. The authors extract essential elements from a horseback archery and reconstruct them in more simple style utilizing fusion of various sensors and movable L-shape immersive screens. The developed system enables ordinary people to enjoy the excitement of a horseback archery. Moreover, the authors import competition between two players into the traditional style of horseback archery so as to improve quality as entertainment.

抄録全体を表示

In this paper, several electronically enhanced board games are described. The motivation of our work conies from a traditional board game, which a group of players enjoy in a face-to-face situation by manipulating pieces on a physical board. With E^2board (Electronically Enhanced Board), that has been devised by the authors, a physical board and a virtual world created by a computer are linked together, and therefore a game itself will be more exciting and immersive. In order to construct inboard, a function of quickly recognizing massive objects which are different from each other is necessary. We use the RFID technology and devise a new method for parallel processing of data transmission. By using E^2board, we have so far developed several applications: not only games but also systems for supporting collaborative work and learning. In this paper, systems called BoardHockey, CatchTheButtefly, BoardFighter, and Symphony-Q are shown.

抄録全体を表示

We analyzed the movement of a dancer by Kansei information processing technique, trying to assume the image that the dancer intends to express. Next, we developed an interactive dance system which supports the dancer's emotional expression using the analyzed data. For evaluating the support effect of transmitting such image expression, we conducted an experiment to show the synchronized image that combines the image and music suitably represented the designed image of the dancer to the experimental subjects. As a result, we confirmed that the appropriate assumption method and synchronized contents would help the dancer to express his/her expected image more easily.

抄録全体を表示

This paper presents a novel method for analysis and synthesis of human actions in movies. This method can capture human actions from movies by fitting an articulated model to the human body image in the movie. This non-invasive method enables us to capture actions even from old movies. Results of capturing actions from real movies are shown. Furthermore, replaying the captured actions from some movies can costar the digital actors in one short animation.

抄録全体を表示

This paper introduces a novel method for interpolation of viewpoints at a soccer match. According to the property of soccer scene, object images are classified into several regions and intermediate view images for each region are generated based on the projective geometry. Then synthesizing them completes new view images of the whole soccer scene. Since weak camera calibration provides sufficient information for the proposed method, it can be applied to a dynamic event in a large space such as a soccer stadium. We give some examples of video images we have generated that make the viewers feel as if they are flying through the real soccer field. This approach will lead to the creation of completely new and enjoyable ways to present and view soccer games.

抄録全体を表示

3D video is the ultimate image medium recording dynamic visual events in the real world as is. Recorded object behaviors can be observed from any viewpoint, because 3D video records the object's full 3D shape, motion, and precise surface properties (i.e. color and texture). In this paper, we present research attainments so far obtained: 1. a PC cluster system for real-time reconstruction of dynamic 3D object behaviors from multi-view video images, 2. an active 3D mesh model for reconstructing the accurate 3D object shape, 3. an algorithm of generating natural-looking texture on the 3D object surface from the multi-view video images, and 4. an editing system for visualizing 3D video with an omni-directional background image using versatile 3D camera works. Experimental results demonstrate the effectiveness of these methods in generating high fidelity and attractive object images from arbitrary viewpoints.

抄録全体を表示

Copying human action accurately is main scheme in entertainment VR area to control and generate a virtual human or cartoon character in a screen. Motion capture system is generally used to generate a crone human in the scene of motion picture or interactive game, however, huge manual operation at post processing is inevitable to generate high quality image. In this paper, especially copying facial action is focused on and high quality and accurate method to generate and copy natural impression of face with semi-interactive process using face image analysis and face image synthesis scheme. Finally, some application systems in entertainment area are introduced.

抄録全体を表示

This paper studies the real-time interactivity in the DVE (Distributed Virtual Environment) with lag. That is, the influence of degradation on real-time interactivity is examined, in terms of the existence of time lag in the DVE shared by the several avatars representing remote users on the network. In this perspective, this paper proposes the method called PFL (Predictive Feedback Loop) in order to compensate the degradation by using the prediction for the current status of the remote avatars. The historical data about the remote avatars and the time stamps in the receiving data are used for prediction with the assumption of stable variation of avatars' status. A simple collaborative task is employed to examine the task performance for experimental evaluation, where the pairs of examinees are forced to work collaboratively together in the peer-to-peer networks with the existence of simulated time lags.

抄録全体を表示

The concept of R-Cubed (Real-time Remote Robotics: R^3) aims to provide a way to telexist anywhere in the world by controlling remote robots over the network. RCML (R-Cubed Manipulation Language) is considered to be a bottom-up approach of the R-Cubed concept. The design of an RCML system utilizes existing infrastructures and devices such as the Internet and PC and, users of the system will be able to use it easily and intuitively. RCML is a language for describing the interface for controlling remote robots in an RCubed concept. This paper proposes a new design of the RCML system called RCML 2.0. In RCML 2.0, we introduced a language RXID 2.0 for defining Graphical User Interface (GUI), which is used for controlling the remote robot into the system. Both RCML 2.0 and RXID 2.0 are XML- based languages. By using XML, expandability and flexibility in implementation are introduced to the RCML system. RXID 2.0 has mechanism for a one-way link to an RCML data structure, and this mechanism provides for the complete separation of the control of the robot and the user interface. We also show the experimental implementation of the RCML 2.0 system which supports PDA and celler phone as a client in addition to ordinary desktop PC client.

抄録全体を表示

A control methodology for master-slave systems using passive force feedback has proposed by the authors. The methodology solves the conventional problems of previously developed master-slave systems with force feedback, such as oscillations, complex structures and complicated control algorithm. In the present paper, a multi-fingered exoskeleton haptic device (master hand) with passive force feedback function is developed. First, the exoskeleton master hand with three fingers (12 degrees of freedom) is designed and produced. Each finger of the master hand consists of a link mechanism with elastic-shaft joints and clutches. The master hand makes fingertip positions and angles of index finger, middle finger and thumb to be measured using link mechanisms. Furthermore, it also enables passive force feedback to an operator by the same link mechanism used for the geometric measurements. Then, a virtual reality system of human hand is constructed using the master hand and the control methodology. Using the system, sensory evaluations are conducted to confirm the usability of the developed master hand and the possibility of the control methodology in virtual reality system. Virtual reality system using the master hand and the methodology is constructed. As a result, the subjects possibly recognize the stiffness of the objects in the virtual environment.Haptic Device

抄録全体を表示

The purposes of this study were to survey the current state of virtual reality(VR) exposure therapy and to examine the tasks needed to refine it by reviewing VR exposure treatment studies during last 1 1 years(1990-2000). As a result of literature search, there were 23 treatment studies since 1994, it was found that VR exposure was the effective treatment for specific phobia such as acrophobia, fear of flying, and for panic disorder, and post traumatic stress disorder. It was pointed out that evoking a sense of presence in the virtual environment was essential for successful VR exposure treatment. It was discussed that personal and virtual environmental factors which affect presence and treatment effect must be revealed for more effective VR exposure.

抄録全体を表示