ITE Technical Report
Online ISSN : 2424-1970
Print ISSN : 1342-6893
ISSN-L : 1342-6893
23.79
Showing 1-21 articles out of 21 articles from the selected issue
  • Type: Cover
    Pages Cover1-
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (14K)
  • Type: Index
    Pages Toc1-
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (83K)
  • Wen ZHONG, Miki HASEYAMA, Hideo KITAJIMA
    Type: Article
    Session ID: BCS99-63
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Active contour models are effective in contour extraction. One difficulty with models the technologies is that much work is needed for reliable parameter acquisition, which is conventionally done by trial-and-error methods. This paper introduces reinforcement learning to determine parameters for specific applications. By modifying the learning stream, our system overcomes a drawback exisiting in the reinforcement learning method that it was not robust. Exploration time is shortened by adding heuristic rules. Experimental results show that our systems work effectively.
    Download PDF (933K)
  • Yoshinori OHWAKI, Masaki HARADA, Toshiaki FUJII, Tadahiko KIMOTO, Masa ...
    Type: Article
    Session ID: BCS99-64
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    The block matching method is used in various scenes as a fundamental technology of the image processing. However, there are much ealculation amount on block matching method, and there is a defect of tdking time. In this paper, we propose high speeding technique of block matching which uses square error as a criterion of the distance between blocks. This paper shows that the indirect decision treatment by the simple calculation sorts out the block and this reduces the block which is calculated square error, and that it can reduce calculation time of the block matching.
    Download PDF (586K)
  • Masaki HARADA, Toshiaki FUJII, Tadahiko KIMOTO, Masayuki TANIMOTO
    Type: Article
    Session ID: BCS99-65
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In the conventional fractal coding schemes, affine transform parameters are calculated. Because the amount of calculation is very large, huge encoding time is required. In this paper, we propose two new fractal image coding schemes to reduce the calculation time. The first one calculation time of parameters, affine transform and rmse by using the maximum amplitude ratio which is a ratio between the maximum amplitude range of range block and that of domain block. This scheme uses a ratio of a variance of range block to that of domain block instead of a scaling parameter. The results of simulations show that the encoding time of the proposed scheme is 30%〜50% smaller than that of conventional schcme.
    Download PDF (538K)
  • Hiroki SUZUKI, Yukiko KENMOCHI, Kazunori KOTANI
    Type: Article
    Session ID: BCS99-66
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In the generating CG images like oil paintings, two processes are necessary. The first is deciding the brush position according to the composition and sketch. The second is generating the brush stroke by using a paint brush, a canvas, and oil colors. We classify the type of brush strokes and make models to shape the brush strokes according to the brush types and the canvas. We show that the characteristics of oil painting brush strokes such as the scratchiness.
    Download PDF (721K)
  • Junko IWAMOTO, Yukiko KENMOCHI, Kazunori KOTANI, Ichiro NAGASAWA
    Type: Article
    Session ID: BCS99-67
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    There are many cultural assets which are desired to being restored and preserved. One of the most damaged assets are wooden statues which have many fractures and wormholes caused by their material properties. Therefore, it is necessary to carry out the urgent restoration. In this paper we focus on the examination and estimation of wormholes We first present the three-dimensional graph representation of wormhole regions, and then extract the graph structures from the X-ray CT images.
    Download PDF (834K)
  • Yasuhiro Himeno, Yukiko Kenmochi, Kazunori Kotani
    Type: Article
    Session ID: BCS99-68
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    There has been some researches for making a snowing scene by using the techniques of computer graphics[1-4]. They have used the simple models of the shapes and the motion of snow to reduce the calculation time, since their research interest was in reducing the calculation time rather than improving realistic images. Conversely, we focus on making more realistic scene of snowing rather than the cost of calculation. First we present the shape model of snow crystals besed on Nakaya-diagram[5]. Next we combine snow crystals to make a snow flake by applying the results of Magono-report[7]. From the shape information of snow flakes, we obtain some parameters for volume rendering which generates an artificial image of snow. Finally we discuss the motion of snow flakes for making snowing scene.
    Download PDF (704K)
  • Yasuyuki KATASE, Masaya MIWA, Shiro HANDA, Shinjiro OSHITA
    Type: Article
    Session ID: BCS99-69
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In case of digital radio communication, the amplification which uses electric power efficiently utilizes even the non-linear area of the amplifier. Then, SER(Symbol Error Rate)performance degrades with distortion of AM-AM and AM-PM. In this paper, the spectral expansion, inter-symbol interference and SER of APSK are considered in the case where non-linear amplification is used. Also, SERs of multiple-symbol differential detection and coherent detection are examined over an AWGN channel and a Rayleigh-fading channel. As a result, it is confirmed that SER performance of multiiple-symbol differential detection is superior to that of coherent detection, when these detection schemes are applied to a typical constellation of 16APSK, the star QAM.
    Download PDF (718K)
  • Euijin Kim, Miki Haseyama, Hideo Kitajima
    Type: Article
    Session ID: BCS99-70
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    We describe how to vote in the parameter space of the Hough transform in a digital image. It has been proven that the calculation cost can be reduced by using the direction and length of the segments. However, the previous algorithm can not represent the value of the correct peaks in the parameter space because the parameter space was defined as a discrete intercept and slope interval. In this paper, we propose an algorithm that uses the association of the intercept and slope parameters and the length of the segments of the digital lines. We show by experimental results that the algorithm can detect the correct value of the peaks in the parameter space.
    Download PDF (586K)
  • Nobuki SAKAMOTO, Kustarto WIDOYO, Terumasa AOKI, Hiroshi YASUDA
    Type: Article
    Session ID: BCS99-71
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    This report describes the Gaze Communication for 'MONJUnoCHIE' videocoference system, which realizes teleconferences with eye-contact, life-size image and proximity to displayed persons, At first we examined what image should be displayed to support eye-contact from any point of view, and we found out that attendees should keep on looking at the camera behind the display independently of the observer's point of view. Based on the result we discussed some image displaying methods. Secondly we examined the accuracy of the recognition of windows in the displayed shared space between 2 remote sites. And we discussed applying the result to MONJUnoCHIE system.
    Download PDF (819K)
  • Yoshimitsu KUROKI, Yoshifumi UESHIGE, Teiji OHTA
    Type: Article
    Session ID: BCS99-72
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In the lossy mode of Joint Photographic Experts Group(JPEG), Discrete Cosign Transform(DCT)coefficients are entropy-coded either Huffman coder or binary arithmetic coder called QM-coder. In the case where QM-coder is employed, DCT coefficients are converted into binary sequences and then context-modeled referring the statistical model. This paper presents that the distributions of AC coefficients are not Gaussian, but Laplacian through Kolmogorov-Smirnov(KS)test statistic. Furthermore we propose a state assignment scheme based on a Laplacian distribution to achieve higher compression performance than JPEG.
    Download PDF (544K)
  • Yoshifumi Uchiyama, Miki Haseyama, Hideo Kitajima
    Type: Article
    Session ID: BCS99-73
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    This paper proposes new parameters of a Hopfield Neural Network(HNN)for edge defection of images. The HNN has been used as a preprocessor of an image restoration HNN. Therefore, edge detection accuracy is closely related with the image restoration performance. The proposed new parameters can make the previous method detect edges more accurately.
    Download PDF (698K)
  • Chikara Ushimarau, Miki Haseyama, Hideo Kitajima
    Type: Article
    Session ID: BCS99-74
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    This paper presents a multi-agent based method to extract the same objects in image sequences. In the proposed method, the objects are divided into line segments which are considered to be agents. Further, the method is based on a multi-agent algorithm of which the utility function is defied with a motion continuity and a shape classification of line segments. By optimizing the utility, the proposed method searches for a suitable line segment which corresponds to the target segment in a different frame. According to the group rationality, the proposed method can correctly provide correspondence among line segments in different frames. Further, the computational cost can be reduced by utilizing the group rationality. The proposed method can also extract occluded objects by tracking utilities in multi frames.
    Download PDF (597K)
  • John GATES, Miki HASEYAMA, Hideo KITAJIMA
    Type: Article
    Session ID: BCS99-75
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    This paper presents a high-accuracy real-time polygon extraction and classification algorithm. The algorithm is capable of extracting both convex and concave polygons from complex images. The algorithm can classify triangles into five distinct classes and can classify quadrilaterals into six distinct classes. The algorithm was tested with a variety of natural and synthetic 256×256 grayscale images and an average frame rate of more than 69 frames / second was obtained using a 450 MHz. Pentium II Processor. Experimental results are presented which demonstrate the high-accuracy of the algorithm.
    Download PDF (774K)
  • Tomohiro Mae, Nobumoto YAMANE, Yoshitaka MORIKAWA
    Type: Article
    Session ID: BCS99-76
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    A 2-stage Vector Quantization(VQ)method by reflection group is proposed to implement large dimensional and high rate VQ. This method is sequential search and direct sum codebook constrained 2-stage VQ that is able to reduce computation and memory amount. And because all of Voronoi regions of the first stage VQ are congruent, quantization performance asymptotically increases toward to Rate-Distortion Bound. In this paper, a performance improvement method is proposed, in which some vectors in the codebook of the second stage VQ are arranged to surface regions on the Voronoi region of the first stage VQ to improve asymptotic property without increasing computation and memory amount.
    Download PDF (493K)
  • Takeshi NOMURA, Yoshitaka MORIKAWA, Nobumoto YAMANE
    Type: Article
    Session ID: BCS99-77
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    The authors previously showed an optimization of allpass polyphase dilter banks via minimizing cell area leads good coding performances for subband coding. No optimizations of bite impulse response filter(FIR)banks, however, have confirmed. So this paper examines that this optimization works well for three FIR filter banks ; quadrature mirror filter(QMF)banks, conjugate quadrature filter(CQF)banks and generalized lapped orthogonal(GenLOT)banks.
    Download PDF (1035K)
  • Hidekazu TAKAHASHI, Nobumoto YAMANE, Yoshitaka MORIKAWA
    Type: Article
    Session ID: BCS99-78
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In this paper, a Wiener Filter(WF)based on the Gaussian mixtured distribution model is adopted to restoration of blured mages with additive noise. In this method, image signals in local blocks are classified into finite number of Gaussian processes and WF's are applied to each process individually to adapt WF to local variation of image statistics. In the case of restoring bured images, detection of the model parameter is significantly suffered with noise and performance of restoration is reduced. To improve detection of the model parameter, a regularization method is adopted. Finally simulation results show the efficiency of the proposed method.
    Download PDF (827K)
  • Youichi KAWAKAMI, Nobumoto YAMANE, Yoshitaka MORIKAWA
    Type: Article
    Session ID: BCS99-79
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In Wiener filter method based on the Gaussian mixture distribution model, image signals in arbitrary shaped blocks are observed to be classified into finite number of Gaussian processes to adapt WF to local variation of image statistics. In this paper, a performance improvement in this method is proposed in the case of large noise variance. In this case, required size of the filter support becomes large and outer block pixels that have different statistics are used as filter support signals. In the proposed method, the outer block pixels are included in the observation to prevent using the outer pixels that have different statistics as filter support signals. A simulation result shows that the filter performance is improved with the proposed method.
    Download PDF (705K)
  • Kennichi OSUMI, Yoshitaka MORIKAWA, Nobumoto YAMANE
    Type: Article
    Session ID: BCS99-80
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    This paper discusses variable shape partitioning method to improve coding performance of fractal image coding. In the quad-tree partitioning method, which is well known as one of variable shape partitioning, we divide detail region into small range blocks to improve quality of reconstructed image. However, since the number of range blocks increases, the rate-distortion characteristics are not necessarily improved. In this paper, we first introduce a partitioning method in which we control blocks such that their sides stand parallel to local directions and the number of pixels in block remains constant, and discuss the selection method to decide the fitting domain blocks. In the last, we carry out coding simulations for our and the existing methods, and show that rate-distortion characteristics of our method outperforms and the quality of reconstruction are much improved for images including many edges.
    Download PDF (1122K)
  • Type: Appendix
    Pages App1-
    Published: December 16, 1999
    Released: June 23, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (86K)
feedback
Top