Reports of the Technical Conference of the Institute of Image Electronics Engineers of Japan
Reports of the 226th Technical Conference of the Institute of Image Electronics Engineers of Japan
Showing 1-20 articles out of 20 articles from the selected issue
Date: July 28, 2006 Location: Muroran Institute of Technology
9:30-10:50 Chair: Hiroshi OCHI, Nihon University
  • - Spectrum · Discrete Cosine Transform · Genetic Algorithm · Lemma -
    Yukio IMAI, Takehiko TOMIKAWA
    Session ID: 06-01-01
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    It is a report of the design of the image information processing by the information technology and the genetic algorithm. The design procedure is enumerated. The image information is resolved to the orthogonalization spectrum image information in the beginning. Next, compression processes the orthogonalization spectrum image information. Finally, the image information system is evaluated. The previous two items are repeated. The best image information system is found.
    Download PDF (203K)
  • Kazuya SASAZAKI, Yukinori SUZUKI, Junji MAEDA
    Session ID: 06-01-02
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    We proposed vector quantization with variable code-vectors for image compression. The variable code-vectors could solve a tradeoff between encoded image quality and a compression rate of images. We developed algorithm to solve the problem using local fractal dimensions. That is vector quantization with variable code-vectors. Result of computational experiments showed the effectiveness of proposed algorithm.
    Download PDF (495K)
  • Hirotaka SUZUKI, Madoka HASEGAWA, Shigeo KATO
    Session ID: 06-01-03
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    We propose a lossless video coding scheme using Markov model coding. In Markov model coding, the number of the contexts increases exponentially depending on the number of reference pixels and the gray levels of the image. In this case, the amount of code of Markov table we have to transmit also increases, Markov model coding is irrealizable. However, in this study, because we take a method to transmit one piece of Markov table per one video image sequence, we can keep an overhead low bit rate. The simulation results show that the entropy of Markov model coding using motion compensation is less than coding rates of JPEG2000.
    Download PDF (474K)
  • Kazuharu YANAGIHARA, Ichiro MATSUDA, Susumu ITOH
    Session ID: 06-01-04
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Motion-compensated prediction based on precise motion vectors, typically with half-pel or quarter-pel accuracy, is commonly used in current video coding schemes. This paper proposes a motion-compensated prediction method which can compensate precise motions using two-dimensional adaptive interpolation filters. In this method, plural number of non-separable two-dimensional interpolation filters are designed for each frame, and an appropriate combination of the filter and a motion vector with integer-pel accuracy is determined in each macroblock composed 16 x 16 pels. Moreover, a shape of the filter mask is adaptively changed to reduce prediction errors with smaller amount of side-information required for filter coefficients. Simulation results indicate that the proposed method provides better coding performance than the conventional methods using motion vectors with fractional-pel accuracy.
    Download PDF (889K)
11:00-12:20 Chair: Tomohiko ARIKAWA, NTT Resonant Inc.
  • Masaki TAKAHASHI, Toshihiko MISU, Mahito FUJII, Nobuyuki YAGI
    Session ID: 06-01-05
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A visualization system for tee shot trajectories in live golf programs is developed. The system extracts and tracks ball regions in ball-following camera images, and measures ball directions. The ball trajectory curve is calculated based on the ball directions and it is superimposed on the ball-following camera image. The system uses only a regular broadcasting camera, and doesn't require any additional camera or operator. Moreover, the system estimates the ball fall timing and fall position by predicting the three dimensional ball position sequentially. Therefore, the pointing operation is semi-automated, and the system can draw the ball trajectory immediately.
    Download PDF (697K)
  • - Utilization of the Regularity of Parallel Cameras -
    Takanori SENOH, Terumasa AOKI, Hiroshi YASUDA, Takuyo KOGURE
    Session ID: 06-01-06
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    For the future television system, multi-view video coding is under the standardization in International Standardization Organization (ISO), which will be used for Free-viewpoint TV or 3D TV. If the camera arrangement is known, a fast and accurate prediction is possible. In this report, assuming the object face direction, the relationship between camera images is derived and utilized for the block displacement and deformation prediction. Preliminary experiments showed bi-directional prediction, variable block size prediction and reference block shape compensation are effective for disparity compensation. The possibility of disparity vector compression is also investigated utilizing the relationship between them.
    Download PDF (765K)
  • Kennosuke YASUHIKO, Sato SAGA
    Session ID: 06-01-07
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    We have already proposed a sketch-based 3-D modeling interface, BlueGrotto, on the basis of a freehand curve identification technique named FSCI-3D. This is an interface with witch a user can input each of solid primitives as a sweeping operation, where a sweep path and a sweep profile are input by freehand drawing in virtual-reality space. However, with BlueGrotto, it is impossible for a user to input complex 3-D objects that are defined by several profiles, such as a ship's hull. This paper proposes to introduce a skinning operation into BlueGrotto and establishes a new user interface with which a user can input multiple profiles by freehand drawing. Experimental results show that a user can input a variety of 3-D objects as sweeping operations and/or skinning operations using the proposed user interface in a consistent freehand drawing manner.
    Download PDF (648K)
  • Naoki KON, Sato SAGA
    Session ID: 06-01-08
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    We have already proposed a front-end processor BlueGrottoFEP which provides a 3-D sketch recognition service to existing CAD systems, and demonstrated that a combination of BlueGrottoFEP and a practical CAD system works as a practical sketch based CAD system. However, in the combination, a user is requested to synchronize contents between BlueGrottoFEP and the CAD system manually, and switching from one to the other becomes a troublesome operation. This paper realizes an automatic synchronizing mechanism in the combination of BlueGrottoFEP and an existing practical CAD system, and shows that the combination can work as a CAD system where a user can freely switch between sketch operations in BlueGrottoFEP and command operations of the practical CAD system.
    Download PDF (711K)
13:10-14:30 Chair: Makoto MATSUKI, NTT Quaris Corp.
  • Kosuke AOKI, Wataru TAKAHASHI, Shizuo NAKANO
    Session ID: 06-01-09
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    When taken outdoor, color information of pictures changes greatly by environmental changes, setting of digital camera, etc. This research reports an effort to improve of color reproduction using color chart currently used to judge the growing state of crops, which will help perform adequate promotion diagnosis using image communication.
    Download PDF (624K)
  • Hiroto YOKOGISHIZAWA, Junji MAEDA
    Session ID: 06-01-10
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    In this paper, we propose the edge-preserving smoothing method as a pre-processing of perceptual image segmentation. The proposed method combines the dynamic morphological filter and the anisotropic diffusion algorithm for smoothing color images. The technique can provide the smoothed color images by removing small features such as fine textures and weak edges that are not necessary for perceptual image segmentation. We experimentally show the effectiveness of the proposed method by comparing it with the conventional methods.
    Download PDF (842K)
  • Masayuki MATSUNAGA, Junji MAEDA
    Session ID: 06-01-11
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    This paper presents the color quantization using a self-organizing map (SOM) and image segmentation. We try to implement perceptual segmentation of color images using only the color quantization and without using the texture features. We propose an improved hierarchical SOM that is a modified version of the Ong's method. The proposed method achieves more stable color quantization results than the conventional method. We also propose to use the pre-processing including the edge preserving smoothing by an anisotropic diffusion that is suitable for reducing small regions and realizing perceptual segmentation.
    Download PDF (615K)
  • Ken-ichi TANAKA
    Session ID: 06-01-12
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    In this article, Simulated-Annealing ( SA ) is introduced into the color halftoning. SA is one of the method of combination optimization problem. SA is the superior method compared with the GA, because of no generation of discontinuity appeares in boundary area, and processing speed is fast. It becomes an excellent expression by limiting the color made a quantum according to the color of the original image when SA is introduced into the color halftoning. Moreover, the generation of the color other than white and the black are lost in the monochrome image.
    Download PDF (593K)
14:40-16:00 Chair: Takanori SENOH, The University of Tokyo
  • Hiroyuki ASANO, Yuji SAKAMOTO
    Session ID: 06-01-13
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    RGB method was commonly used to express colors in a grating-type rainbow hologram. We propose a new color expression method for holograms that is using two appropriate spectra in a white reconstruction light. This method can realize large fringe pattern area and high luminance nearly about one point two-times comparing with conventional methods theoretically. The method is applied to display color images by grating-type rainbow hologram on CDR, successfully. We discuss the proposed method and report experimental results.
    Download PDF (194K)
  • Hirokazu TOMONAGA, Katsuyuki SHINOHARA
    Session ID: 06-01-14
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    With access opportunity of image contents increasing, effective image retrieval is required. Although there are many works of Content Based Image Retrieval (CBIR), a lot of them refer to the whole of database to compute their degree of similarity. Then, we aim to applying hashing method, which search the part of database for rapidly image retrieval, to CBIR. In this report, before applying hashing method to CBIR, we propose a feature compression method which transform Hue-histogram one of color feature from multidimensional features into one-dimension feature and check the classification accuracy of this method. From the results of experiments, the maximum accuracy in this method was 0.71. As this method base, we improve the classification accuracy of compressed color feature, and then we apply hashing method to CBIR in the future.
    Download PDF (300K)
  • Masaru KUBOTA, Katsuyuki SHINOHARA
    Session ID: 06-01-15
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    The system with which required information is retrieved efficiently is required now when a huge quantity of information overflows. Also in the field of a picture, it is the same and needs to classify a picture beforehand for the improvement in efficiency of search. So, this paper considers the classification technique of a picture based on correlation between pictures. It normalized to each axis of the amount of the picture features, and dignity of each axis was made equal. As a result of measuring correlation between pictures based on these, a normalization front and after normalization, the clear difference appeared in correlation between pictures. Therefore, a threshold value required in order to classify a similar picture can be set up easily.
    Download PDF (324K)
  • Jun WATANABE, Madoka HASEGAWA, Shigeo KATO
    Session ID: 06-01-16
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Studies of digital museum are making progress in various research institutes and universities. It is very important to leave great art works and cultural heritage for the future generation. These works are digitized, archived and stored in the digital museum. These digitized works can be shown without spatial restriction and the original works can be kept out of aged deterioration. However, the digital museum has a problem that no one can verify integrity of the digitized art works because of the nature of digital data. In this paper, a digital watermarking method for the digitized art works of 3D objects to detect error or tampering is proposed. We treat a 3D object as a group of still images. In our method, watermarking and adaptive coding are combined and the watermark is embedded in the images by using redundancy of choice of coding parameter. Signature of each image is embedded as a watermark in another image, so that the images are integrated each other. The special feature of our method is that the images never degrade by the embedding of the watermark.
    Download PDF (349K)
16:10-17:30 Chair: Ken-Ichi TANAKA, Meiji University
  • Kuniaki YAJIMA, Norio TAYAMA
    Session ID: 06-01-17
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    We had proposed the WMR method. The images can be restructured from few number of projection data at high speed. The algorithm is composed of the wavelet sampling function and the SVD. Therefore, the reconstruction became difficult because of an increase in the number of sampling points. To decrease the number of sampling points in the reconstruction area, the triangular sampling was applied. In addition, we applied the multi wavelet sampling and the block reconstruction method. As a result, the accuracy was improved and the calculation time was shortened.
    Download PDF (166K)
  • Shunya HATAKEYAMA, Madoka HASEGAWA, Shigeo KATO
    Session ID: 06-01-18
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    In this paper, we propose a new method for representing gray level image by two-tone. In our proposed method, the whole area is divided into the blocks composed of 2 × 2 pixels, and determined a total value of pixels in each block. Secondly, the value is divided into the blocks composed of the maximum pixel value. The quotient means the number of white dots. The position of the white dots is decided by the order of pixel value. The rest is saved for the next process. Then, neighbor 2 × 2 blocks are merged and the above-mentioned distribution process is recursively applied. Thus, the position of white dots is progressively determined, that is to say, this algorithm works as bottom-up process. In addition, we consider improvement of our proposed method. As a result, our proposed method outperforms the Multistage Division Quantization method ( MDQ ) in the gradation region.
    Download PDF (253K)
  • Yusuke CHIHARA, Hiroshi KANEKO
    Session ID: 06-01-19
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    The conventional fractal dimension is scalar-valued and cannot contain sufficient image information for image analysis. For example, spatial information such as the directionality and the configurative information are not represented by the fractal dimension. Overcoming these defects, in this paper, a new type of fractal feature, which we call the directional fractal dimension, is proposed and applied to image analysis. The feature is a generalization of box counting dimension and three fractal dimensions along x, y, z axis are calculated respectively, which contains directional information of the figure. The experiment verifies that the fractal feature is very useful for image analysis and classification.
    Download PDF (528K)
  • Hirotaka KADOSHIMA, Yuji SAKAMOTO
    Session ID: 06-01-20
    Published: July 28, 2006
    Released: January 31, 2007
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Existing shape morphing has been transformed by using a linear interpolation. Therefore compared with real transformation of the real material, the transformation using existing method is unnatural. Moreover, a linear interpolation needs corresponding of the points beforehand that is troublesome work. In this text, it proposes a new shape morphing method that solves these problems. We move and transform a target object by moving the ball which was made for inside of object by Bubble Mesh. As a result, each part of the object achieve an organic transformation by complexly and respectively moving, and a natural transformation of the materials such as clay and water was obtained. Moreover, the work of corresponding of the point became unnecessary, and morphing became possible easily.
    Download PDF (198K)
feedback
Top