Toward the ubiquitous network society, the cooperation with enterprise, university, municipality and the citizens in the region is paid attention to be as promotion power of IT. In 2000, we established the regional IT consortium and have been developing a digital welfare map as the practicing research. The digital welfare map can provide welfare information such as the obstacle and life information on the Internet, by which the senior citizen and physically handicapped persons can go out safe. In this paper, first outline of IT consortium was introduced. Next, the system structure and technical studies concerning system design of the digital welfare map model system were shown. From the results of the regional proof experiment which volunteers participated, the effectiveness of this model system was confirmed. Finally, future subjects concerning promotion of IT in the region were discussed.

View full abstract

This paper proposes a Multicast-like delivery tree construction protocol and a video merging algorithm of the Active Video Delivery (AVD) mechanism that can be utilized for delivering a multiple video digest. It considers the problems that exist in the current state of IP Multicast networks and addresses issues regarding the efficiency of delivering streaming video content. The multiple video digest system provides users with an initial multiple video preview upon their requests, instead of requiring a direct download of large video files. With such a system in place, two problems can be solved: (i) the shortening of time required to retrieve videos from large-scale video archives; and (ii) helping users to choose the desired content efficiently.
  The results in our analytic model and simulations indicate that with a properly chosen AVD Sharing stream regeneration threshold, our proposed mechanism significantly reduces the resources, such as the total network cost, the average link load, and the video server bandwidth requirement, which are required to distribute the requested video content. In addition, results based on simulations confirm that the AVD tree can achieve the same efficiency as that of conventional IP Multicast approaches.

View full abstract

It is very important to control the bit-rate for video communication under a limited capacity of communication channel like wireless communication system because of large amount of data. Although there is a function to change the bit-rate in the quantization process of the standard video coding scheme itself, the controllable range is restricted and insufficient for more local controlling. From this point of view, we propose an adaptive pre-filter to improve the bit-rate in video compression, which preserves visually significant portion of the original image and smoothes only insignificant portion actively. Furthermore, we propose three kinds of temporal extension of this scheme and verify their performance on compression ratio and visual image quality.

View full abstract

We have developed an advanced vector quantization (VQ) algorithm for high quality real-time image transfer system. The error optimizing adaptive resolution (EAR) method can realize much superior compression performance than the JPEG and the JPEG2000. In the compression of the XGA (1024×768 pixels) still-images including text, for instance, there exists an overwhelming performance difference of 2 to 40dB in compressed image quality. In order to increase the performance of image compression by Vector Quantization (VQ), we propose a systematic codebook design method without using learning sequences for 4×4 and 2×2 pixel blocks. According to the method, the codebook can be applied to all kinds of images and exhibits equivalent compression performance to the specific codebooks created individually by conventional learning method using corresponding images.
  Furthermore, we have also developed a real-time wireless transfer system for still images based on EAR-VQ. The system can transmit a XGA still image within 2.1 sec.

View full abstract

In this paper, a novel error concealment algorithm is proposed for the MPEG-4 video decoding. The MPEG-4 standard adopts a set of error resilient tools to improve video robustness in error-prone environments such as wireless communication. Error concealment scheme can improve image quality for images with the residual errors which are not corrected although applying the error resilient tools. However, since error concealment technique is not specified in the MPEG-4, there are strong demands to develop novel high performance error concealment technique. The problem for conventional error concealment algorithms is high computational complexity or a low performance for video sequences with high motion. The proposed algorithm adopts a hybrid scheme, which can achieve consistent performance for various video sequences at various error rates, and can reduce computational complexity using the temporal and spatial correlation of macroblocks. This hybrid scheme is composed of two techniques; the modified DMVE technique and adaptive spatial searching technique. The modified DMVE technique reduces the computational cost by using the temporal and spatial correlation of motion vectors, and adaptive spatial searching technique can conceal high frequency components by searching for similar pixels within a frame. The proposed algorithm is constructed on the basis of the MPEG-4 software decoder, and the experimental results demonstrate that our approach provides acceptable performance both subjectively and objectively at various bit error rates.

View full abstract

In the recent networked society, 24-hour internet and e-mail access is available via mobile terminals such as cellular phones and PDAs, and in the near future, internet access will become indispensable in our daily life. Therefore, it is very important to improve the readability of text on low resolution displays. However, since the readability of each character can only be evaluated by subjective methods, designing fonts for specific devices requires a long development period. The readability is dependent on “center and center of gravity”, “balance of character parts”, “counter”, and “stroke”. In this paper, considering the center of gravity of characters through experiments, we reveal the correlation between subjective and objective evaluations. Furthermore, it is considered that the result of our experiments can complement the development tool for “LCFONT”, which is distinctive fonts optimized to LCDs, by the efficient objective evaluations in conjunction with the original subjective evaluations, and can reduce greatly the development period of new fonts.

View full abstract

The encoded speech for digital transmission systems contains semantic information and singular information. Therefore, a speech or speaker recognition system can be realized by using only the encoded speech. In this paper, we propose a speaker verification method based on a speech coding scheme in the digital transmission systems. The proposed method utilizes CELP (Code Excited Linear Prediction) parameters which are used in speech coding schemes for mobile communication systems or IP networks, and verifies a speaker only with the encoded speech. The merits of the proposed method are as follows; (1) Speaker verification is easily realized in the current mobile terminals or network systems by adding a little function. (2) Since CELP parameters contain a speaker's characteristics of articulation, text-independent speaker verification is realized. The reliability of the proposed method is discussed with some simulation results.

View full abstract

This paper describes a proposal of an effective anchor setting method for video hypermedia using fusion of automatic tracking based on motion picture analysis and manual editing. We can take these advantages, (1) improving the accuracy of tracking using bi-directional tracking, (2) editing anchor effective by deleting intermediate frames by using proposed anchor setting method. And we implemented prototype anchor setting software based on this method and made experiments for evaluation of sport video. As a result, we evaluated results of experiments and proved its effectiveness.

View full abstract

In this paper, we propose video editing system which is based on Server-Client system over broadband network. This system is composed of video editing server and client PC. One of the most important things is that the client PC needs only general Web browser. This system can reduce the cost for generation and editing the video contents. After we discuss the change of circulation of video contents over the telecommunication network, we design and implement our proposed system. Through the experiment on editing time, the effectiveness of proposed system can be confirmed.

View full abstract

It is a fundamental and important problem to detect Weak Perspective Projection invariant shapes from an image. Although there was the conventional method of using the tangent information on a figure outline, since the straight line figure was undetectable, we proposed the method of extracting an eternal axis for weak perspective projection conversion using the figure center of gravity. This method is applicable not only to curvilinear form but straight line form and the form containing both straight line and curve. However, For the method based on the result of center of gravity calculation, if a deficit occurs in the figure, the center of gravity will move and detection will become difficult. This paper proposes a new method without center of gravity calculation and it can detect a deficit figure.

View full abstract

In this paper, we suggested a flat type of two observers-tracking parallax stereoscopic display system with three parallax images. There are two display methods can be used for this system. The one rearranges three parallax images in to a compound image in order and the two viewers can always observe stereoscopic images. The another prepares several series of parallax images and displays three parallax images correspond observers' tracking position. In this case, one observer can observe the stereoscopic images with full motion parallax and another observer can observe the stereoscopic images with partial motion parallax.

View full abstract

The method to restore warped page by using 3-dimension shape computation of the end of book is already proposed. However, it cannot be used for twist, and is difficult to rectify small characters correctly because the 3-dimensional form of the page is not calculated. In several experiments, we discovered that the form of the page of a book is a developable surface and the curved surface consisted of straight line elements. This paper proposes a method of restoring page by extracting the straight line elements and rearranging the straight line elements.

View full abstract

This paper describes comprehensive architecture of digital rights management system (DRMS) for content delivery services over the network. Uniqueness of the architecture is to properly manage the rights of content throughout entire distribution stages with robustness of content protection and flexibility in license management. This is also independent on the format of content and the delivery methods such as file downloading or streaming. This capability is established by introduction of the layered rights processing stack, secured content packing technology, and PKI (Public Key Infrastructure) based transaction scheme. This paper describes an implementation of DRMS into digital video and audio distribution system over the Internet, and we discuss validation of the proposed architecture based on digital video delivery.

View full abstract

There is an attempt to realize indoor space such as museums by 3-dimensional computer graphics virtually. In this paper, we propose a simple modeling method for 3-dimensional indoor space in consideration of disposed objects, to reproduce rooms or underground shopping areas virtually by low cost, using digital camera images. In this method, indoor space and objects are modeled by a set of hexahedrons connected each other. Major objects within the indoor space, such as pillars, are modeled during the course of the modeling of indoor space. The geometry data and the texture data of indoor space and objects can be generated by digital camera images of inner surface areas of each hexahedron in the semiautomate process.
  We reproduced an underground shopping area with objects such as pillars and so on, and we proved that our proposed method is practical.

View full abstract

In this paper, it is shown that there is useful method for improving PSNR of lossy Hadamard transform. This method cancelled mean component of the rounding errors which is generated at rounding process in transform, with offset added at rounding process in inverse transform.
  As a result of improving PSNR of lossy Hadamard transform, partial of input data space is reversibly transformed. Input data space can be divided into two partial spaces. One is the space in which total sum of data is even. Another is the space in which total sum of data is odd. There are many vectors from which offset matrix can be generated such that one of tow partial spaces is reversibly transformed.
  We propose new lossless 2D-Hadamard Transform which exchanges two offset matrix based on that the sum of data is odd or even. It is shown that entropy of the lossless transformed coefficients is scarcely affected by selection of offset matrix.

View full abstract

In orthopedic medicine, many operations are conducted to replace a bone with an implant. In the case of femoral stem fixation in total hip arthroplasty, the initial fixation is important. There are two concepts to consider for an ideal fixation: One common concept is that a wider contact area is better for fixation because lower stress must be better for biological tissues. Another concept is that the fixation force should be distributed to avoid the stress shielding of bone. Optimal design of joint prostheses has been generally investigated under these concepts. In this study, the 3-D FEM was used to map the stress distribution on the contact area of a virtual model of a bone tissue and prosthetic stems. The CT imaging of DICOM data was used to construct of a FF stem and an IMC stem models. Furthermore various mathematical model that simulating the contact mechanism between the materials and bone tissues were constructed. As a result, the wider the apparent contact area, the greater the fluctuation. The contact area can be limited by IMC stem fixation. It was concluded that the general concept unconditionally recommending a larger fixation site area for joint prostheses, should be revised.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract