This paper describes a real-time 3-dimensional head tracking for an indoor wearable navigation system using augmented reality, which provides navigation information as virtual objects overlaid on the real scene. We propose a hybrid tracking with a pair of stereo-camera and an orientation sensor, which localizes a 3-dimensional position and orientation of a user's head. Discretely located infrared markers in the workspace are used as fiducial markers. Through some experiments, it is shown that this tracking system has reasonable accuracy enough to provide navigation information.

抄録全体を表示

In this paper, a novel personal telerobotic system named "telecommunicator" is introduced. It has a movable head with a video camera and a simple arm. Both the head and the arm can be telecontrolled from the local site through the Internet, using a wireless communication. As a result, the person on the local site can communicate with people around the telecommunicator on the remote site. Two different forms of the telecommunicator are proposed; a wearable one and a mobile one. To confirm application capacity of the wearable telecommunicator, a prototype T1 has been developed as a research platform. Furthermore, a pointing function is discussed with a human property. In addition, a compensation technique against motions of human body is introduced.

抄録全体を表示

Galvanic vestibular stimulation (GVS) induces a sensation of virtual acceleration as vestibular information. GVS is expected as a wearable interface because it does not need a large device like a motion platform. In particular, GVS with alternating current (AC) can influence vision which mainly consists of torsional component. The purpose of this research was to evaluate quantitatively the effects of GVS on visual perception, and to infer its cause. In order to investigate these issues, we conducted psychophysical experiments during GVS with AC that consisted of three different types of image presentation method: Image fixed on a spatial coordinate system, a head coordinate system and a retina coordinate system. It is suggested that the visual motion perception induced by GVS is mediated by eye movements.

抄録全体を表示

In this paper, we propose a display-based position and orientation measurement system. It consists of a display device and photo sensors that can track multiple receiver positions and orientations simultaneously. A fiducial marker is projected onto the center photo sensor of the receiver. If the movement causes the photo sensors to detect the fiducial marker's position and orientation, the projector moves the fiducial marker to the center of the receiver again canceling the offset between the receiver and the fiducial marker. Many types of projection systems, including displays such as LCDs and desktop monitors can use this tracking technology.

抄録全体を表示

ウェアラブルコンピュータを活用することで、利用者の生活の質を改善できると期待されているものの、まだ必ずしも一般に広く受け入れられるには至っていない。本論文ではその原因を解明するため、マズローによる人間の基本要求階層を応用することを試みる。本理論を活用することで、一般人に受容されうるウェアラブルコンピュータの特性を解明できる、というのが筆者らの仮説である。そこから導かれるのは人間の基本的要求を満足するウェアラブルコンピュータのコンセプトであり、筆者らはそれをサイバークローズと呼ぶことにする。第一次のアンケート調査を行ない、このコンセプトにより基本要求階層のうち生理レベルと安全性レベルの要求を満たしうることを確認した。しかし社会的帰属レベルについては必ずしも明確な結果を得られなかった。この調査結果をより明確にし、かつ仮説の妥当性を検証するため、プロトタイプ機を製作し実験を行っている。特に初対面での自己紹介に応用する局面について、この実験を紹介した。

抄録全体を表示

In recently proposed saccade-based information displays, the observer cannot see an afterimage without moving the eyes appropriately and therefore often cannot see the afterimage correctly. We propose a new display system based on a concise saccade detection technique and the perceptual features of eye movement. In addition, with a system consisting of a wearable sensor (saccade detector) and a ubiquitous display (saccade-based display), we propose an application where many viewers can be accommodated simultaneously (Several viewers can view an image simultaneously (timesharing) with one saccade-based display.) and independently (Several viewers can view different images with one saccade-based display.). In experiments, we compared character-recognition rates to confirm that the detection of the saccade improves the visibility of the saccade-based display. The results indicate that the recognition rate is clearly improved.

抄録全体を表示

In this paper, we describe a novel function of ActiveCube that realizes intuitive interaction with multimedia contents. First we realize a new method that reduces the host PC's load by distributed control between the host PC and cubes. Secondly, we reduce unnecessary packets in the network between the host PC and cubes using autonomy and encapsulation of each cube. Then we conducted experiments to evaluate the performance of the proposed system. Finally, we developed an example application for children and observed children's actual actions.

抄録全体を表示

This paper describes the application of location oriented mobile system in National Science Museum Tokyo. Developed system was introduced as a real world role-playing game. About 20,000 children have experienced our game. In this paper, we describe briefly about system and argue about interaction and evaluation of game-like content for museum guide application by analyzing participants' behavior. We provided two different types of content. First content is a stamp rally type game story, which rule is simple enough to understand even if user experiences our application for the first time. Second content is designed to process a brief comment about an exhibit when user comes close to it just like museum visitor's companion does.

抄録全体を表示

A small and lightweight wearable equipment has been developed for experience capturing. It can capture human-human and human-object interactions with user's person's view video image and sound annotated by the interacting user/object IDs by means of infrared LED ID tracker. These data are used for sharing experiences among people. This paper presents the design factor through prototyping of usable wearable experience capturing system and experimental usability evaluation of the prototype.

抄録全体を表示

K-eXplorer is an ubiquitous real-world interface consisting of a K-tai (smart phone) and K-stations that operate between people and objects. By using K-eXplorer the user is able to access, link and manage information of real-world objects tagged with an RFID tag/QR Code, stored in a remote database through a K-tai and several K-stations installed in commodities in the workspace. K-stations have multiple I/O ports that input data from a K-tai, real-world objects, the server and the user to transfer it to the other. We have developed a prototype system of the K-eXplorer which implements three K-stations (a tray type, a desk light type and a bag type) that the user is able to refer and link to the objects with. The usability of K-stations is evaluated and the result indicated the favorable potential of the system.

抄録全体を表示

This paper describes a Functionally-Distributed (FD) hand tracking method for hand-gesture-based wearable visual interfaces which is an extension of the Distributed Monte Carlo (DMC) tracking method. The method provides coarse but rapid hand tracking results with the lowest possible number of samples on the wearable side, and can reduce latency which causes a decline in usability and performance of gesture-based interfaces. In addition, the method can provide the adaptive tracking mechanism by using the sufficient number of samples and the hand-color modeling on the infrastructure side. This paper also introduces several applications of the hand-gesture-based wearable visual interfaces implemented on our wearable systems.

抄録全体を表示

As to individual equalization of differences in HRTFs (Head-Related Transfer Functions) between a dummy-head and a listener for a binaural system, there is a constraint condition that an equalization function must be measured for the listener in the original sound field. In order to resolve the condition, a method to replace the function measured in the original sound field by that measured in another sound field (an equalization sound field) is proposed in this article. Since profiles of the two functions are similar, the original function is approximated by applying ARMA (Auto-Regressive Moving-Average) model to the function measured in an equalization sound field. Experimental results showed that the proposed method was effective for all subjects when the numbers of poles/zeros were 60/60 in ARMA model.

抄録全体を表示

Virtual reality (VR) technologies have recently been applied to traditional arts and industrial area. In painting field, there are several researches to make realistic brushes and pens for drawing and calligraphy. However, a realistic method with a haptic device still has not been implemented yet. In this paper, we propose a virtual calligraphic brush model for a haptic device to enable a natural drawing. We observed force variation of a writing brush, and make our model based on a spring-dumper model. Area of grounded brush, compression power, and friction force were considered. The implemented VR calligraphy system is evaluated by experiments, and it is shown that the proposed model can achieve realistic calligraphy feelings.

抄録全体を表示

The number of science museums that have virtual reality systems is increasing. This paper proposes an exhibition method using immersive projection technology by a researcher in a science museum. The researcher uses virtual reality contents as the exhibition contents and explains the contents to audiences. That is to say, the method enables communication between researchers and audiences. We exhibited contents that are related to genome science three times in the science museum according to our proposal method. We report this exhibitions and results of questionnaires from audiences. This paper also describes about a visualization method of audiences' operation log in a virtual environment. The museum exhibition can get feedback from this visualization results.

抄録全体を表示

This paper proposes a volume masking framework for interactive editing and physics-based modeling on volume images reconstructed from voxel data. The proposed framework handles volume masks colored by voxels without editing voxel data directly. Using the masks, cut-and-paste, move and deformation are represented as volume images in real time. As practical application, we developed volume masking and interaction software. The performance and effectiveness are confirmed in preliminary use for interactive editing and elastic object modeling on several volume images.

抄録全体を表示

Recently, a number of methods for 3-D modeling from images have been developed. However, the accuracy of a reconstructed model depends on camera positions and postures with which the images are obtained. In most of conventional methods, some skills for adequately controlling the camera movement are needed for users to obtain a good 3-D model. In this study, we propose an interactive 3-D modeling interface in which special skills are not required. This interface consists of "indication of camera movement" and "preview of reconstruction result." In experiments, we have confirmed this interface makes it possible for users to obtain a good 3-D model easily without a special training.

抄録全体を表示

In order to implement high-definition sound localization system such as auditory display, acquisition of individual Head-Related Transfer Functions (HRTFs) without direct measurement is essential. As the fundamental study of the precise prediction of HRTFs, relation between the HRTFs and the anthropometric parameters was investigated for each subject by means of canonical correlation analysis. The results suggested that not only the head-related parameters but also the ones concerning the subject's torso have significant effect to his/her Interaural Time Difference (ITD). A method to estimate the azimuthal variation of the ITD for each subject from his/her anthropometric parameters was then proposed in this paper. Accuracy of the estimation with the proposed method was evaluated via a hearing test. The results showed that the adjustment of the ITDs for each subject with the proposed estimation method significantly improve his/her azimuthal localization accuracy.

抄録全体を表示

This paper presents a method for simulating human motion with tissue shape deformation based on an elastic skeletal muscle model. Shapes of tissue models are decided based on real measurement data such as CT. In the model, motion is rendered by muscle contraction such as bending knees, and the bending motion and shapes of soft tissues are generated as a result of dynamic simulation. The soft tissues are modeled by our original elastic model which has advantages of low computational cost and stability in large deformation. The result of tissue shape deformation is transparently visualized as a real-time animation.

抄録全体を表示