The Journal of the Institute of Image Electronics Engineers of Japan
Online ISSN : 1348-0316
Print ISSN : 0285-9831
ISSN-L : 0285-9831
Volume 39, Issue 2
Displaying 1-10 of 10 articles from this issue
Contributed Papers
  • Kunihiko TAKANO, Kota YAMAKOSHI, Kazushige SUNAHARA, Genta KAWADA, Kok ...
    2010 Volume 39 Issue 2 Pages 134-139
    Published: March 25, 2010
    Released on J-STAGE: August 25, 2011
    JOURNAL FREE ACCESS
    In the holographic 3D display using mist screen, large 3D images can be observed with wide viewing angle. However, in the previous methods, it was difficult to observe the sides of the projected 3D images. In this paper, we studied a process to expand the displaying volume of the projected 3D images by preparing a jetting system improving the characteristics of the mist screen and by combining the reconstructing waves reflected from two SLM's. As this result, by adopting our system, 3D images are seen to become observable with wider viewing angle including sides than before.
    Download PDF (1360K)
  • Motoki MORITANI, Fumihiko SAITOH
    2010 Volume 39 Issue 2 Pages 140-147
    Published: March 25, 2010
    Released on J-STAGE: August 25, 2011
    JOURNAL FREE ACCESS
    This paper proposes a template matching method, (ELM: Extreme gray-levels Location in local area Matching) which is based on location of extreme gray-level pixels in local area. In the method, the extreme gray-level pixels location in a local area is registered to every pixel in a template image. The objective image area is searched from the matching ratio, which is obtained from sum of difference between objective image and template image whose location of maximum and minimum gray-level pixels in local area. The experimental results show that the proposed method had equal or better robustness to the inferior factors of objective images in comparison with the typical two conventional template matching methods.
    Download PDF (2367K)
  • Kousuke IMAMURA, Hideo HASHIMOTO
    2010 Volume 39 Issue 2 Pages 148-155
    Published: March 25, 2010
    Released on J-STAGE: August 25, 2011
    JOURNAL FREE ACCESS
    Extraction of semantic moving objects from video sequence is an indispensable tool that benefits newly developed image processing schemes. In the present paper, we propose a new technique that extracts shapes of moving objects by combining SNAKES and Watershed algorithm. The proposed method comprises two steps. In the first step, SNAKES extract contours of moving objects as a result of the minimization of an energy function. In the second step, the conditional Watershed algorithm extracts contours of moving objects from a topographical surface including a new function term. This function term is introduced to improve the estimated contours considering boundaries of moving objects obtained by SNAKES. The efficiency of the proposed approach in moving object extraction is demonstrated through computer simulations.
    Download PDF (1593K)
  • Souta TSUCHIYA, Fumihiko SAITOH
    2010 Volume 39 Issue 2 Pages 156-162
    Published: March 25, 2010
    Released on J-STAGE: August 25, 2011
    JOURNAL FREE ACCESS
    We propose the automatic distinction technique of natural object image and artificial object image. This method uses a statistical distinction of the distribution concerning the edge direction of a natural object image and an artificial object image. The proposed method can classify them by using statistical information without a parameter. This method make weighted edge direction histogram with the edge direction and strength of the input image. A frequency ingredient of this weighted edge direction histogram can judge whether the target object image is a natural object image or an artificial object image, by comparing the distance between the target object image and natural object image, and the distance between the target object image and artificial object image. The experimental results show that the high distinction precision by the proposed method.
    Download PDF (2579K)
  • Yingdi XIE, Jun OHYA
    2010 Volume 39 Issue 2 Pages 163-174
    Published: March 25, 2010
    Released on J-STAGE: August 25, 2011
    JOURNAL FREE ACCESS
    This paper proposes a modified RANSAC based method that can detect multiple independently moving objects from the video sequences acquired by active stereo cameras, and can estimate the egomotion of the stereo cameras. We first compute 3D optical flows from consecutive frames, where dense 3D optical flows, which are needed for the subsequent egomotion estimation based on the modified RANSAC, are obtained by the process proposed in this paper. From the dense flows, three 3D optical flows are randomly selected for computing an egomotion (model). The validity of that egomotion model is checked using the weight obtained by calculating a consistency between that model and each of the dense flows. If it turns out that the estimated egomotion model is valid, the label that identifies this model is assigned to the inliers, which are consistent with this model, and then the labeled inliers are deleted. The random choice of three flows and its subsequent processes are repeated till one of the termination conditions is met. Using the inliers present in the static background, the real egomotion is computed. Experiments using synthesized and real stereo sequences demonstrate the effectiveness of the proposed method.
    Download PDF (1343K)
  • Alireza AHRARY, Jian ZHANG, Sei-ichiro KAMATA
    2010 Volume 39 Issue 2 Pages 175-184
    Published: March 25, 2010
    Released on J-STAGE: August 25, 2011
    JOURNAL FREE ACCESS
    Signature verification is a challenging task, because only a small set of genuine samples can be acquired and usually no forgeries are available in real application. In this paper, we propose a novel approach based on Hilbert scanning patterns and Gaussian mixture models for automatic on-line signature verification. Our system is composed of a similarity measure based on Hilbert scanning patterns and a simplified Gaussian mixture model for decision-level evaluation. To be practical, we introduce specific simplification ways for constructing a model and its training method. The system is compared to other state-of-the-art systems based on the results of the First International Signature Verification Competition (SVC 2004). Experiments are conducted to verify the effectiveness of our system.
    Download PDF (516K)
  • Chao-Ning CHAN, Shunichiro OE
    2010 Volume 39 Issue 2 Pages 185-194
    Published: March 25, 2010
    Released on J-STAGE: August 25, 2011
    JOURNAL FREE ACCESS
    A new active eye-tracking system has been developed using the frequency image processing method in order to improve the effects of the central coordination of eyes when investigated under various lighting conditions. In this paper, we have proposed the use of quad PTZ cameras to track the face and eye; further, we use image compensation with the frequency image processing method, which can compute the central position of the eye and face detection improvements under low illumination levels. In all the tasks, the frequency image processing method was able to separate the pupil and the white of the eye, and further, feature detection under various light environments was also improved. The user's gaze points are calculated on the basis of the position of the eyes and face and calibrated by cognitive calibration methods. Our system provides better usability for the user who can now operate the system without any restrictions. The eye-tracking system is capable of calculating eye gaze in real-time to provide a responsive interaction.
    Download PDF (2257K)
  • Ryo OHTERA, Takahiko HORIUCHI, Shoji TOMINAGA
    2010 Volume 39 Issue 2 Pages 195-203
    Published: March 25, 2010
    Released on J-STAGE: August 25, 2011
    JOURNAL FREE ACCESS
    Captured images with a digital camera are sometimes different from the appearance of the human vision. The Retinex model is a well known algorithm for image quality improvement with considering visual characteristics. The Retinex can reproduce an image by removing the influence of the scene illuminant. However, the complicated processing is required to reproduce an appropriate image. In this paper, at first, we construct a high-speed system for image quality improvement by devising both the improvement of the Retinex algorithm and the implementation of hardware. Next, we show that the proposed system can reproduce almost the same quality images as conventional MSR's by subjective evaluations. Then, we also show that the calculation time of the proposed system is greatly improved. Finally, the display method and the display size of the improvement region for concerning convenience are examined by subjective evaluations.
    Download PDF (896K)
  • Susumu MARUNO, Taro IMAGAWA
    2010 Volume 39 Issue 2 Pages 204-211
    Published: March 25, 2010
    Released on J-STAGE: August 25, 2011
    JOURNAL FREE ACCESS
    We have already proposed Temporal Pattern Recognition Network with Adaptive Segmentation of Quantizer Neuron Architecture (TASQA) and φ-s transformation of shapes for object recognition using camera images and we proved their robust object recognition with rotation free. Hardware implementation of TASQA, including ASQA network which is the main recognition part of TASQA, is necessary for practical application. We newly developed specialized LSI for ASQA network called ASQA chip and we realized extremely high speed and robust object shape recognition system. In this paper, the structure and the ability of ASQA chip and the general structure of the object shape recognition system are described.
    Download PDF (1198K)
Tutorial
feedback
Top