ITE Technical Report
Online ISSN : 2424-1970
Print ISSN : 1342-6893
ISSN-L : 1342-6893
23.9
Showing 1-18 articles out of 18 articles from the selected issue
  • Type: Cover
    Pages Cover1-
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (17K)
  • Type: Index
    Pages Toc1-
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (99K)
  • Jong-Il Park, Yuichi Iwadate
    Type: Article
    Session ID: IDY99-85/HIR99-20/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Toward efficient video annotation, we propose an estimation of global camera parameters from digital video. The method is to directly utilize motion vectors contained in a video stream. The estimation is based on an iterated extended Kalman filter with outlier rejection. Experimental results with both synthetic and real data show that the method is well suited for extracting global camera motion information from digital video.
    Download PDF (1041K)
  • Jong-Il Park, Masahiro Chiba, Yuichi Iwadate
    Type: Article
    Session ID: IDY99-86/HIR99-21/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In this paper, an efficient method of generating an arbitrary view from multiple cameras. The method is based on first extracting sharp depth map and then utiliting the map to generate an arbitrary view. Toward realtime view generation, efficient implementations based on careful insight into epipolar geometry are proposed and demonstrated.
    Download PDF (761K)
  • Naoki KISARA, Takaaki HASEGAWA
    Type: Article
    Session ID: IDY99-87/HIR99-22/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    This paper describes extraction of three-dimensional information from monocular image sequences of a forward moving camera. Considering monocular moving image sequences as stereo images, and searching corresponding points, three-dimensional information that is extracted, and tracking methods of corresponding points, are discussed. Moreover, the measurement error affected by the moving distance of the camera, searching methods of corresponding points, compensation of the optical axis's offset, are discussed. At last, from experiment results, it is confirmed that three-dimensional information is extracted by this method.
    Download PDF (984K)
  • Manabu Kato, Yuji Sakamoto
    Type: Article
    Session ID: IDY99-88/HIR99-23/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    To realize the surgery simulation using virtual reality, it is necessary to define a deformation and cutting method of the soft objects such as muscle, organ and etc. Therefore, we propose a method modeling the objects by a mass-spring model generated from 3-demensional volume data of X-ray CT, MRI and etc, and define a deformation and cutting by the mass-spring model. In this method, it is possible to generate a mass-spring model from volume data directly by related a voxel and mass point as one to one. Deformational operation is done using a force, move, fix, and cut operators. It is able to represent various actions by combination of these deformational operators.
    Download PDF (789K)
  • Hiroki Nitta, Wataru Mitsuhasi
    Type: Article
    Session ID: IDY99-89/HIR99-24/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In this paper, we report a system to track an object moving parallel and vertical to the lines of sight of observing camera. Dilational and rotational motion in the cartesian image coordinates are mapped to translations in the log-polar coordinates, and thus easily estimated by using a general method. Although, translations in the cartesian coordinates are much complicated ion the log-polar coordinates, they have a remarkable characteristic with periodicity in the mapped domain. Thus the translational motion can be efficiently estimated in the log-polar cooridnates. A tracking system is constructed with a CCD camera of which pan, tilt and zoom are externally controlled. This system can track a moving object with processing speed of 12frames/second without extra hardwares.
    Download PDF (1136K)
  • Akihiko SAWADA, Wataru MITSUHASHI
    Type: Article
    Session ID: IDY99-90/HIR99-25/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    A method of feature detection for human face images is proposed. One of the remarkable features on human faces is "eye", which has a symetrical structure in image intensities. Some symmetrical transforms on image intensities are presented for detecting face features such as eye, nose or mouth. Image areas centered around the detected features are mapped to log-polar coordinates and then processed by auto-correlation to obtain rotation and scale invariance. Finally the auto-correlation is classified in the feature space by primary component analysis. Experimental results are also presented.
    Download PDF (983K)
  • Masayuki BABA, Yoshiaki KATO
    Type: Article
    Session ID: IDY99-91/HIR99-26/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In Japan, satellite and terrestrial digital broadcasting will start from the year 2000. Various data with various features are multiplexed in digital broadcasting. We propose a new multiplexing control method. It can be used both for integrated multiplexing that deals with various media, and for statistical multiplexing that controls each media data rate to give proper bandwidth. In this multiplexing control method, data priority for multiplexing is decided based on delay time to multiplexing, possibility of abandon, and ability of rate control.
    Download PDF (1295K)
  • Hidenori Yamamoto, Hidehiko Iwasa, Haruo Takemura, Naokazu Yokoya
    Type: Article
    Session ID: IDY99-92/HIR99-27/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Recently, importance of an image retrieval system which can retrieve intended images from huge image databases has been increasing. In past years, many image retrieval systems have been proposed. Color histogram is one of popular image features and it has been exploited in many image retrieval systems. A color histogram captures the global color distribution of an image, but it lacks information of spatial relationship among image pixels, i.e., two images with the same color histogram may have completely different layout. This paper proposes an image retrieval system using the layout of colors in an image. The proposed system divides an image into some reqions based on spatial color information in the image and computes a color histogram from each region. Experimental results show that our method can retrieve images that users intended in comparison with a single color histogram method.
    Download PDF (884K)
  • Takahiro Ushiba, Kiyoshi Kiyokawa, Haruo Takemura, Naokazu Yokoya
    Type: Article
    Session ID: IDY99-93/HIR99-28/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Polygonal models, which are often used in realtime 3-D computer graphics such as video games and virtual reality applications, are desired to have small number of polygons as possible, as long as they preserve original shape and appearence. Conventional polygonal simplification methods do not utilize texture images that ere often mapped onto polygonal models. In this paper, we propose a novel technique of view-dependent polygonal simplification that utilizes a texture image that contains appearances of an original model. We also describe an empirical study on the effectiveness of the proposed method.
    Download PDF (1324K)
  • Kentaro Kichiyoshi, Hidehiko Iwasa, Haruo Takemura, Naokazu Yokoya
    Type: Article
    Session ID: IDY99-94/HIR99-29/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    This paper describes a method for supporting query-based data mining. Recently, data mining has attracted much attention as a methodology of discovering useful knowledge from large databases. Data mining is the task that discovers a description expressing common regularities among data. It is a general method of data mining to find a description expressing a subset of data from the database automatically, but such an approach has a problem that the method obtain the results that are not desired by users. Therfore, we employ a query-based data mining scheme in which a user repeats the verification and modification of user's hypothesis, and propose a method of supporting the query-based data mining process by Visualizing databases. We also show the case study on query-based data mining by using the newly developed prototype system.
    Download PDF (1114K)
  • Nagato NARITA, Masaru KANAZAWA, Ichiro YUYAMA
    Type: Article
    Session ID: IDY99-95/HIR99-30/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    The researches regarding psychological factors have been carried out in order to develop a television with higher sensation of reality. The impressions that Hi-Vision and 3-D Hi-Vision(with polarized glasses)sequences gave to observers were analyzed by showing fourteen sequences with the SD(Semantic Differential)technique. The opinion scores were analyzed by factor analysis, and eight psychological factors were obtained for each above formats. In addition, it was examined whether the relationship between the opinion scores obtained using seven-grade quality scale and the factor scores could be represented by multiple regression models. Consequently, it was clarified that the regression model applied to the scores for Hi-Vision sequences and did not to those for 3-D Hi-Vision sequences.
    Download PDF (596K)
  • Katsuhiko Masuda, Makoto Omodani, Yasusuke Takahashi
    Type: Article
    Session ID: IDY99-96/HIR99-31/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    The purpose of this study is to clarity quantitatively the difference of thinking faculty between two conditions using softcopy and using hardcopy. We have examined two working conditions -display medium in horizontal state and vertical state- for both medium of LCD as a softcopy and printing paper as a hardcopy. A series of tests have been carried out by measuring calculation speed and correct answer rate to simple calculation tasks using five members of testees. The results have shown clear differences in the correct answer rate. The speed of calculations in vertical condition are faster than that of horizontal condition. In addition, in horizontal states, the speed is faster when using hardcopy than using LCD, while they are almost the same in vertical state. It is suggested that the advantage of hardcopy originates in its horizontal working condition.
    Download PDF (704K)
  • Megumi Takezawa, Hiroyuki Honda, Miki Haseyama, Hideo Kitajima
    Type: Article
    Session ID: IDY99-97/HIR99-32/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    This paper proposes a method to quantize IFS parameters accurately. The conventional method quantizes them to the nearest values. However the method does not necessarily provide high-quality reconstructed images. Therefore, a new quantization method searches for the optimal IFS-parameter set by using a genetic algorithm to obtain the high-quality ones.
    Download PDF (1004K)
  • Takeshi Mutou, Miki Haseyama, Hideo Kitajima
    Type: Article
    Session ID: IDY99-98/HIR99-33/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    This paper proposes a method which determines correspondence of feature points in two different images by using an elastic model. The previous method cannot do it accurately, when the users do not know which object each feature point exists in. The proposed method can acquire suitable correspondence, by comparison of elastic energies of expression by the proposed elastic model.
    Download PDF (529K)
  • Chikara Ushimaru, Miki Haseyama, Hideo Kitajima
    Type: Article
    Session ID: IDY99-99/HIR99-34/MI
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    This paper presents a method to extract the same closed smooth curves in different frames. Since the proposed method is based on not only the Process-Grammar but also localities of fragments of curves, it can detect the same curves even modified by rotation and magnification. The previous method cannot achieve such detection, because it is based on only the similarity of their motion vectors ; the motion vectors of a rotated or magnified curve are different from the vectors of the original one.
    Download PDF (527K)
  • Type: Appendix
    Pages App1-
    Published: February 03, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (71K)
feedback
Top