We have attempted a conversion from stereoscopic HDTV to multi-viewpoint imaging. From LR images we extracted a parallax by phase-correlation method and then produced the data of parallax by allocating it to pixels. Based on the data, we then produced a camera-interpolated image. To assess the quality of the produced moving image, we have also developed a system that shows the image on several multi-view displays.

View full abstract

We are researching technology for archiving performances of Japanese traditional arts such as "Noh" and "Kabuki". The paper describes a new texture mapping method for the 3D model's surface. Using this method, we have developed a system for real-time and continuous display of high-resolution arbitrary viewpoint images of performances. Finally, we performed an experiment on real-time and continuous display of high-resolution arbitrary viewpoint images of "Noh," which is a Japanese traditional dramatic art.

View full abstract

We have proposed "Free Viewpoint VoD system" which reconstructs free viewpoint images using images from one viewpoint and depth maps. This paper proposes a method for transmission of depth information, which extracts the exact background region using background mask in order to reduce the required amount of data without spoiling the accuracy of reconstructed images.

View full abstract

Holographic stereogram(HS) is useful as the method of 3D-TV. So we at first researched HS by an optical technique which is useful as a basic consideration, then we consider HS by a computer technique.

View full abstract

It is a goal that Holographic 3D image transmission using Digital SSTV system is achieved. JPEG2000 is used in Digital SSTV technique. In this paper, the quality of images recovered by a CGH through JPEG2000 is investigated. Applicability of JPEG2000 to the transmission of 3D holographic images is also studied. As this result, JPEG2000 is seen to be applicable to realize a transmission of 3D holographic images.

View full abstract

In this report, the results of computer simulation for image reconstruction from Fourier transform and Fresnel CGHs were presented. The simulation for these CGHs was carried out by applying numerical integration to the Fresnel Kirchhoff diffraction formula.

View full abstract

3D videos are reconstructed from multiple-view images and are represented in a set of static 3D sequences. 3D videos are not suitable for transmission or storage because of their large data size. In this paper, we propose a temporal redundancy compression method with extended block matching algorithm. In addition, we demonstrate some preliminary results of the extended block matching.

View full abstract

In this paper, two methods for temporal 3D video segmentation are proposed, both of which adopt histograms to reveal the motion of 3D object. One utilizes the point distance histograms (briefly PDH) and the other utilizes the histograms in spherical coordinate system (briefly SCH). Our experimental results show both of methods are effective but we also present the performance comparison of these two methods.

View full abstract

The frames of a panned background video sequence photographed at arbitrary speed are reorganized in order to synchronize with the movement of a CG model produced, independently. In order to achieve the synchronization of each frame of both, we propose how to get the panning, tilting and rolling parameters.

View full abstract

We pour easy animation patterns for Light Emitting Diode signs, and make new feeling signboards which always display trademarks. We also consider a policy which does not spoil profits either of a company nor a town.

View full abstract

We make a country precious image and examine how to distribute these effectively together with there local people. And we also consider how to profit by the image media for this area.

View full abstract

For the color design on an outdoor industrial product, it is requested to the CG creating their models not only their shapes but also the match of color to the background photograph. We propose how to photograph and calibrate to acquire all skylight luminance distribution according to the (E_R, E_G, E_B) signal.

View full abstract

The technique of molecular graphics is useful for viewing of the three-dimensional structure of DNAs and proteins. Therefore, molecular graphics programs based on the graphics libraries such as OpenGL and Mesa were developed using computer languages corresponding to these libraries.

View full abstract

We introduced physics simulation to VR(Virtual Reality) world, and constructed the environment like real world. We displayed both VR world and some seconds' future VR world with another physical simulator for prediction. By the suggestion, we displayed future path of table tennis ball on immersive multi-screen display "CABIN".

View full abstract

The interactive television will be available providing variable services in response to an individual demand by using personal data. In order to realize such services, the viewer's personal data should be kept secure and sent to proper receiver reflecting user's will. We report a study for broadcasting services using user information and its protection method.

View full abstract

Emerging IP-based broadcast service is expected to provide new values by using uploaded user information. For privacy protection, we propose the secure communication protocol of user information applying the digitally signed document sanitizing scheme. A prototype system demonstrates that users can check their profiles with feeling of security.

View full abstract

Secure interactive transmission of user information makes it feasible to provide services that reflect viewer information on communications networks. For viewers to feel confident providing information, broadcasters that receive information must inspect and manage it in a secure manner. In this paper, we investigated a database management method for protecting against leakage of user information. In this method, even if the information in the database is leaked, users cannot be specified by the data in the database. We also investigated a database inspection method in which it is possible for items that can be inspected to be limited dynamically at each terminal by the database administrator.

View full abstract

We propose a new contents broadcast system which has copyright protection mechanism and enables each user to copy contents to his personal terminals. Thus we can realize controlled personal use of broadcast contents.

View full abstract

The spread of broadband IP networks is making high-quality video distribution service a reality. However, some technical problems should be solved to deliver programs nationwide with quality and reliability equivalent to that of current broadcast services using radio wave. This study illustrates high level requirements of realizing technology of high quality and dependable IP broadcasting service.

View full abstract

Streaming contents of TV programs over IP network may not be received just as broadcasting corporation intended. It is a result of network trouble like packet losses. This is one of the problems to be solved to realize TV broadcasting over IP network. This paper describes an experimental system which keeps broadcasting the information of programs even if network trouble occurs.

View full abstract

The authors study on a method of detecting audio delay that is required for alignment adjustment of the audio signals in extracting audio features for automatic audio quality monitoring. The proposed method is based on the peak position detection of moving-averaged audio signals. Computer simulations show that alignment adjustment is possible by sending only 12bytes/sec peak position information from both sender and receiver side.

View full abstract

We developed H.264 real-time encoding system with HD full resolution(1920*1080) in 30FPS, and decoding system. The document describe about outline of system with technical description

View full abstract

This paper describes the improvement of rate control technique embedded in the reference software of H.264/AVC Joint Model(JM). We propose an adaptive control of the variance of the quantization parameter(QP) and bit assignment for GOP layer rate control.

View full abstract

A pre-processing filter for improvement of the picture quality of AVC/H.264 encoder is proposed. It eliminates high-frequency unimportant components in temporal moving area and spatial diagonal frequency energy. It also has a function of boost middle-frequency of vertical and horizontal direction. It preserves details of image without bitrate increasing.

View full abstract

H.264/AVC supports various block sizes of motion compensation.(from 16x16 down to 4x4). But, there is an effective block size and not effective block size in SDTV picture. The result of verifying the effective combination of motion compensation block size is reported in this paper.

View full abstract

In this paper, we present the development of an experimental encoder for mobile digital terrestrial broadcasting service. Our encoder system is implemented on Windows PC as a software, and is fully conforming to the standardization of ARIB.

View full abstract

High compression encoding scheme plays a very important role in transmitting super high definition videos under the low bit rate. H.264 FR-Ext is applicable for such high definitions, but motion compensated prediction is conducted by the macroblock whose size is limited independent of the image resolution. From this perspective, extending macroblock size is expected to achieve higher coding performance especially for super high definition video. Therefore, a performance improvement scheme by extending available macroblock size for H.264 FR-Ext is proposed in this paper. The effectiveness of the proposed method is evaluated by several coding experiments.

View full abstract

The prototype of scent television system using datacasting was developed. BML coded scent data is sent to the receiver and the receiver provides the scent data to the diffuser synchronizing with the image.

View full abstract

We have proposed a novel audio watermarking technology, which features nearly lossless embedding, robustness against lossy data compression or analogue conversion, and enables contactless asynchronous detection of embedded watermarks through speaker and microphone devices. Using this audio watermark technology, we propose an application for information broadcasting services, which features distributing digital information through traditional analogue broadcasting or recording infrastructures.

View full abstract

Digital terrestrial television broadcast service for mobile reception ("1 segment broadcasting") in Japan, to be expected to launch in early 2006, will enable "ubiquitous" TV viewing for viewers. Furthermore, it is expected that new kind of high interactive TV services will be introduced via 1 segment data broadcasting. This paper describes an overview of prototype CM-synchronized data broadcasting contents and results of technical verification and open experiments using these contents.

View full abstract

This paper describes an overview of AdapTV services. The AdapTV is a concept of services that present contents in various forms depending on a viewer's context, such as circumstances, abilities and devices. This paper also describes an application of the AdapTV to data broadcasting services.

View full abstract

In this paper, we will present the system requirement for approvable delay time from the request to receiving time of the required content while driving a car. We will also explain some basic parameters such as appropriate information rate for one period of transmission.

View full abstract

We are studying to clarify the condition that can compress and reproduce high-speed moving images with high-resolution and high-fidelity from the viewpoint of viewers with superior eye movement characteristics. Here we introduce the apparatus developed for this purpose and discuss the results about the smooth pursuit eye movements while viewing high-speed moving images.

View full abstract

3D positions of a pupil was measured using stereo wide cameras. A narrow camera was directed to the pupil position and detected the centers of the pupil and corneal reflection. The eye-gaze point calibration method executed by only looking at the narrow camera and a point on the PC screen was proposed.

View full abstract

3D positions of pupils and nostrils were detected using infrared light sources and two stereo cameras. Nostril detection in the image was supported by our robust pupil detection technology. The experimental results showed that head direction detection was possible for a horizontal large angle of the head, ±45 deg, by estimating the undetected nostril position from the metrics data obtained from initial calibration procedure.

View full abstract

We are aiming at development of the high quality moving image. For the purpose, we paid attention to eye movement of sport players who have excellent kinetic vision. We analyzed characteristic of their kinetic vision and eye movement.

View full abstract

We investigated the effect of text readability of a display on eye-movement behavior. Saccade amplitude and fixation duration were incorporated as the eye-movement parameters to understand how affected by different readability factors. When text readability was lower due to lack of luminance contrast, fixation duration was longer and saccade amplitude was smaller. The lower the text readability due to high percentage of kanji, saccade amplitude was smaller, but fixation duration was not changes.

View full abstract

We conducted experiments to describe relationships between viewers' interests to TV programs and their gaze information. We noticed on the so-called zapping as the experimental environment that viewers tend to exhibit their interests on their gaze information. As a result of the ANOVA tests, there were significant relationships between them on some particular experimental conditions.

View full abstract

In this research, the SPL of the auditory stimulus perceived equivalent to the visual stimulus presented in a wide range of the visual field was investigated. Results showed that the equivalent SPL to the visual stimulus in the peripheral visual field is smaller than that in the central visual field.

View full abstract

Pictures taken by a camera with a shutter look better than those without it, especially in moving areas, since blared pictures are sharpened. This has been explained from "apparent motion", i.e., a mystery of a human vision system. But, it is a phenomena related to spatio-temporal sampling. In this paper, the improvement of picture quality mentioned above is analyzed and made clear from this standpoint of view.

View full abstract

We investigated whether the symptom of visually induced motion sickness changed depending on physical display sizes with identical visual angles. Our results showed that a larger physical display size made the symptom of nausea severer when field of view is relatively large (42 deg).

View full abstract

The strength of sickness induced by observations of the two different movie stimuli were compared. The one movie consisted of succession of blank fields and motion scenes. In another movie, in addition, every motion scene included abrupt changes of directions. Observers of the latter reported stronger sickness than observers of the former.

View full abstract

It is indicated that some range of frequency and duration of movement effectively induces the motion sickness and reducing the acceleration of the movement weakens the motion sickness.

View full abstract