All issues
Predecessor
26.83
Displaying 1-20 of 20 articles from this issue
- |<
- <
- 1
- >
- >|
-
Article type: Cover
Pages Cover1-
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (12K) -
Article type: Index
Pages Toc1-
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (77K) -
Purim Na Bangchang, Toshiaki Fujii, Masayuki TanimotoArticle type: Article
Session ID: BCS2002-78
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSWe proposed a new realtime video ray-space data acquisition and rendering system named the "Free Viewpoint Television". With this system, the user can freely change the viewpoint of any dynamic real-world scene in realtime. The basic idea is based on the ray-space method which the arbitrary photo-realistic view can be generated from a collection of real images. For our prototype system, 16 CCD cameras are used to form one dimensional camera array. The interpolation is based on the adaptive filtering ray-space data interpolation. Although only consumer-class hardware is used, the result is good in term of both image quality and rendering speed.View full abstractDownload PDF (574K) -
Mehrdad PANAHPOUR TEHRANI, Toshiaki FUJII, Masayuki TANIMOTOArticle type: Article
Session ID: BCS2002-79
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSTo speed up the processing time, we decentralized the processing tasks to generate new view. In proposed methods, each sensor node processes part of corresponding search algorithm based on MSE minimization to generate the interpolated image with local communication between sensor nodes. In conventional method, all the processing tasks to generate a new view were centralized in central node. We proposed two methods, which are Fully Image Shared Decentralized Processing (FIS-DP), and Partially Image Shared Decentralized Processing (PIS-DP). In FIS-DP, full image transformation should be done between neighbor camera sensor nodes, but the correspondence check is performed in different block intervals. In PIS-DP, the correspondence check is performed in part of images in each node by sharing part of captured images. The experimental results of FIS-DP, and PIS-DP show that we can increase the processing speed of the whole network to generate a new view in compare with Centralized Processing (CP) method. The PIS-DP method is recommended according to its better processing speed and communication efficiency in compare with other methods (FIS-DP, and CP). Using PIS-DP, will let us to perform real-time free view generation in camera sensor network.View full abstractDownload PDF (471K) -
Pao Sriprasertsuk, Wataru Kameyama, Masahito KawamoriArticle type: Article
Session ID: BCS2002-80
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIn the Dresent time, the fusion of the computer world, the communication world and the broadcast world is progressing. In addition, the development of multimedia technology standards and advanced technologies are changing the present audio-visual world (TV beyond TV), that can make various new service models. For example. TV-Anytime Forum is defining service models in which users can explore and acquire TV content distributed over a variety of sources. Moreover, users wm be able to create their own digital entertainment packages exchange and share information among users and redistribute contents among devices, systems or users. To make this real, metadata exchange in distributed environment is a key technology. Besides, Web services seem to have become the next important thing in computing today, because web services are representing an important evolutionary step in building distributed applications. Therefore in this paper, implementation of metadata exchange over IP networks using SOAP and Web Services are described and some experimental results are also mentioned as well as its evaluation.View full abstractDownload PDF (775K) -
Jinhong SHEN, Seiya MIYAZAKI, Teruamsa AOKI, Hiroshi YASUDAArticle type: Article
Session ID: BCS2002-81
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThe aim of our project is to develop an easy-to-use tool that provides an integrated environment for dealing with the complete process of creating digital film. To make the creating process automated, we propose to employ a verbal screenplay as input form. The system we designed can understand the script through a parser, then a virtual director and a virtual cinematographer automatically translate it into a relevant motion picture with various visual effects like real image, three-dimension (3D) animation, or augmented reality. In this paper, "video" stands for all kinds of clips of these visual presentations. On account of the ability to extract suitable video clips from digital video web library and present movie production on web, video data in our system is encoded in XML and tracked by the MPEG-7 standard. This system is an extension of DMP [Seiya and others, 2002].View full abstractDownload PDF (461K) -
Daisuke Semiya, Toshiaki FUJII, Masayuki TANIMOTOArticle type: Article
Session ID: BCS2002-82
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSFractal image coding is weak in errors. When an error occurs in one block, the error propagates to the portion which refers to this block in the decoding process. In this paper, we propose a new coding method for error resilience by sending the image twice with the same amount of data as the single image. This is realized by using double-sized range blocks and a new criterion to determine the domain block. 2-3dB improvement of SNR is achieved in the reconstructed image when an error occurs, while keeping nearly the same SNR when an error doesn't occur.View full abstractDownload PDF (753K) -
Masayuki NAKAMURA, Yuying YANG, Hideki YAMAZAKI, Tomihiko UCHIKAWA, Ka ...Article type: Article
Session ID: BCS2002-83
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSAll cable modems used in CATV use the same upstream and downstream frequencies to establish bi-directional comuunication with their central equipment. The upstream frequency band is usually suffered from noises which causes severe damages to the upstream communication. We think of a method to detect the places where those noises get into coaxial cable of CATV at the central system, using the downstream signals. Then we also developed a tapoff system which has some by-pass switches or filters to cut off subscriber lines which carry the noises to improve the vulnerability of the upsteam frequeny band of CATV.View full abstractDownload PDF (550K) -
Hiroyuki KOBAYASHI, Kisyo TAKAMATSU, Kazunori SHIMAMURAArticle type: Article
Session ID: BCS2002-84
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSWith spreading a broadband network, the number of IP packets is increasing in a backbone network, and will exceed the performance of relay routers in the future. So, the authors studied at the Packet Assembly technology that assembles and reassembles some IP packets on the edge routers of backbone, and mitigates the processor load concerning packet routing of backbone network routers. The author implemented the packet assembly technology by the Data Driven Processor with asynchronous pipelines. This paper is reported extracted problems and the solution.View full abstractDownload PDF (481K) -
Fumitaka SAITO, Hiroyuki YASHIMA, Takuya NAKAMURA, Kazuhiko WAKAMORI, ...Article type: Article
Session ID: BCS2002-85
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThe performance of the atmospheric optical subcarrier multiplex (SCM) systems is evaluated. Three binary PSK (BPSK) signals are multiplexed and modulated by laser diode (LD) operated at nonlinear region. Which causes intermodulation (IM) distortion to degrade system performance. The Signal-to-Noise ratio (SNR) of the received BPSK signal is derived considering shot noise, thermal noise of avalanche photo diode (APD) and second and third order IM. The optimum amplitude of the BPSK giving maximum SNR is shown.View full abstractDownload PDF (574K) -
Soshi MIYASHITA, Junji KAWASAKI, Hirokazu OKAMOTO, Taizo IIJIMAArticle type: Article
Session ID: BCS2002-86
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSWhen we see a binary black and white image, it appears, to our visual sense, to become clearer because of a pseudo halftone. We have investigated and presented a visual model for human visual information processing, which is executed through external world, retina and finally brain. We proposed an objective evaluation method of various modulation images using the above visual model. The Purpose of this paper is to compare PSNR, WSNR, average filter and measure of approximation using the visual model. We made extensive experiments for five kinds of modulation method- ordered dither method, least mean error, pulse density four division, simple two value, random dither- in which each of the images we evaluated both subjectivity and objectivity. The results of a series of experiments show that ranking by subjective evaluation dose not coincide with PSNR, WSNR, and average filter, but coincides with a measure of approximation using the visual model.View full abstractDownload PDF (779K) -
Tetsuo HAOKA, Takashi MASUKO, Takao KOBAYASHIArticle type: Article
Session ID: BCS2002-87
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThis paper describes a method for generating 3D computer graphics animation of hand-gesture using a parameter generation algorithm based on hidden Markov model (HMM).The purpose of this study is to generate motion of gestures renresented by a label sequence. Each label represents a basic motion pattern of the hand, which is modeled by an HMM. When modeling a basic motion pattern by HMM, gesture data, which are parameter sequences of a physical model of the hand, recorded using motion-capturing are used as training samples. Then, given a label sequence, an HMM is composed by concatenating HMMs in the order according to the label sequence, and then a gesture is generated from the composed HMM in a maximum-likelihood sense and put into a computer graphics animation. Smoothness of the synthetic gesture can be achieved by using statistics of static and dynamic features modeled by HMMs. An experimental result shows the effectiveness of this synthesis method.View full abstractDownload PDF (762K) -
Michiharu MAEDAArticle type: Article
Session ID: BCS2002-88
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIn this study, an algorithm of self-organizing maps for image restoration is presented. Self-organizing maps have been hitherto studied for the ordering process and the convergence phase of weight vectors. As another approach of self-organizing maps, a novel algorithm of image restoration is proposed. The present algorithm creates a map containing one unit for each pixel. Utilizing pixel values as input, the image inference is carried out by self-organizing maps. Then, an updating function with a threshold according to the difference between the input value and the inferred pixel is introduced, so as not to respond to a noisy input sensitively. Therefore, the inference of original image proceeds appropriately since any pixel is influenced by surrounding pixels corresponding to the neighboring setting. In the restoration process, the effect of the initial threshold and the initial neighborhood on accuracy is examined. Experimental results are presented in order to show that the present method is effective in quality.View full abstractDownload PDF (1447K) -
Shuichi MAKI, Nobumoto YAMANE, Yoshitaka MORIKAWAArticle type: Article
Session ID: BCS2002-89
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIn the vector quantization (VQ) method, a huge number of computations and memory amounts prevent to realize it in high vector dimensions and high coding rate. Various method have been proposed to ravel out the problem, and unified region-kaleidoscope VQ (UR-KVQ) was proposed as one of them. In high vector dimensions, UR-KVQ is able to design a codebook whose parformance is comarable to the optimum Nearest Nighbor VQ (NNVQ). In this papar, we propose a design method to decide initial codebook of UR-KVQ by rejection using design method for entropy constrained VQ(ECVQ). It is shown that, the codebooks designed by the proposed method have almost the same performance as the NNVQ with small codebook-size, in rate about 3bits per sample in 16-dimensions, and also, lattice-points have minimal norm to origin in oputimum E_8 lattice and related sub-optimum codebook are include in designed codebook.View full abstractDownload PDF (665K) -
Junji TAMURA, Nobumoto YAMANE, Yoshitaka MORIKAWA, Takuichiro WATANABEArticle type: Article
Session ID: BCS2002-90
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIn this paper, a method to detect knitting ball-shaped defects in textile using statistical image processing is proposed, in order to realize automatic visual check. In detecting such defects, a priori knowledge about property of the object image should be used, but in underlying problem, it is difficult to obtain sufficient knowledge about statistical property of the defect image. In the proposed method, we design a statistical model about the nominal image without defects, and use likelihood ratio relative to the nominal model as valuation for the detection, wide sense stationary Gaussian mixture distribution model as the nominal model and normalized DCT AC-coefficient vector on image block as the characteristics vector of the model. In the case that the nominal model is designed using many images that have the same texture as an object image, a simulation result shows that the valuation for detection obtained by the proposed method corresponds well to the visibility of the defects.View full abstractDownload PDF (1213K) -
Kenta MATSUMURA, Masakazu MORIMOTO, Kensaku FUJIIArticle type: Article
Session ID: BCS2002-91
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSTo improve subjective image quality, preferred ROI (Region of Interest) coding has widely been studied. However, in mobile multimedia communication systems, it is very expensive to extract ROI from each video frames. In this paper, we employ MPEG-4 video object coding to realize preferred ROI coding, ROI object is encoded al high-quality to improve subjective video quality and background region is encoded at low-quality to improve coding efficiency. Here, we use ROI tracking method and simplified ROI shape coding to improve coding efficiency and reduce tracking difficulty. Simulation results show that block coding of ROI shape can improve coding efficiency, and results of subjective validation test show that it does not disturb reconstructed video quality.View full abstractDownload PDF (1501K) -
Kazuya NAKAMURA, Yukiko KENMOCHI, Kazunori KOTANIArticle type: Article
Session ID: BCS2002-92
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSEickeler et al. have proposed a face recognition system for compressed images such as DCT coefficients of JPEG. It is pointed out, however, that recognition rates become low when compression ratio are high. Because encoding errors of JPEG2000-compressed images are smaller than those of JPEG-compressed images, we propose a face recognition system for JPEG2000-compressed images which are obtained by wavelet transform. We also show some experimental results comparing with those of Eickeler's system.View full abstractDownload PDF (797K) -
Atsuyuki KURATA, Tomoharu ISHIKAWA, Makoto MIYAHARAArticle type: Article
Session ID: BCS2002-93
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSReproduced colors of electric image devices such as by CRT (LCD,PDP) are different from colors and textures of real objects. New idea is based on the knowledge of color objects in nature world that may be having been keeping their own colors long years. We have adjusted only the value of display gamma and dark level. We have tried to compensate the distortion of total color reproduction system including input and output devices. We have successfully reproduced close colors and textures to objects than traditional color reproduction method.View full abstractDownload PDF (871K) -
Article type: Appendix
Pages App1-
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (74K) -
Article type: Appendix
Pages App2-
Published: December 17, 2002
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (74K)
- |<
- <
- 1
- >
- >|