ITE Technical Report
Online ISSN : 2424-1970
Print ISSN : 1342-6893
ISSN-L : 1342-6893
32.40
Displaying 1-15 of 15 articles from this issue
  • Article type: Cover
    Pages Cover1-
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (14K)
  • Article type: Index
    Pages Toc1-
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (61K)
  • Motomasa TOMIDA, Takanobu TANIMOTO, Satoru ODO, Kiyoshi HOSHINO
    Article type: Article
    Session ID: IDY2008-83/HI2008-10
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Previous systems for human hand posture estimation have adopted clustered multi-layer database with narrowing of search space by its past estimation results. But once an estimated result at a time is out of the search space, the system can't find out a true value. Our system therefore has adopted non-clustered database with narrowing of search space, rather, coarse search at the first stage with some aspects of hand images, and accurate search at the second stage with HLAC. The experimental results showed that averaged estimation error is -4.51 degrees, and the candidates for accurate search at the second stage are reduced from 20,192 to 85.3, which indicates our system realizes the high accuracy and processing speed as previous system without using the past results.
    Download PDF (565K)
  • [in Japanese], [in Japanese]
    Article type: Article
    Session ID: IDY2008-84/HI2008-10
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
  • Hayato FURUKAWA, Kiyoshi HOSHINO
    Article type: Article
    Session ID: IDY2008-85/HI2008-10
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    When the self-collision for the robot hand that had a multi finger and multi degree of freedom was evaded, it was necessary to enlarge the interval of the finger judged that evasion was necessary in the system by making a past finger rest so that what where the finger was made to rest by inertia in an arbitrary place may be the difficult situation, and achieve the collision evasion when the finger operation was fast etc. mutually (hereafter, collision judgment area). However, it becomes impossible to be near the finger when the collision judgment area is enlarged and to operate and the operation of the robot hand will be limited. Therefore, in a small collision judgment area, authors developed the system that did the collision evasion to the finger by avoiding and operating as a technique for achieving the collision evasion. It explains the system as follows.
    Download PDF (422K)
  • Hitoshi HORIUCHI, Satoru KANEKO, Koichi OZAKI, Mie SATO, Masao KASUGA
    Article type: Article
    Session ID: IDY2008-86/HI2008-10
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Recently, 3D shape recognition and depth recognition methods by using monocular camera system are required. These are many studies about 3D shape recognition and depth recognition based on geometric information in short range. In this paper, we aimed to develop a method for depth recognition that is able to recognize depth in a wide sphere. We focused on aerial perspective, a type of depth perception. We considered method for depth recognition based on aerial perspective. The effectivity of our depth recognition method in several meters to kilometers is showed by our experimental results.
    Download PDF (713K)
  • Naoya SUZUKI, Yoshifumi TAKAHASHI, Hisashi OGURO, Sumei GUO, Mie SATO, ...
    Article type: Article
    Session ID: IDY2008-87/HI2008-10
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Binocular disparity is very important in perceiving the distance and we know that the perceptual distances are changed by adjusting binocular disparity. In this study, we research influence on combination of the display size of movie and perceived distance related to adjusting the binocular disparity based on KANSEI evaluation. We experiment in measures the impression of the movie when the combination of the display size and the perceived distance is changed by polarization technique. As a result, we examine the effects on dependency of the viewing condition of the movie as shown by evaluation words impression and correlation.
    Download PDF (534K)
  • Toru Fujiwara, Takuya Kura
    Article type: Article
    Session ID: IDY2008-88/HI2008-10
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Tree-drawing (Baum) test, which is a projective psychological test, is useful to investigate mental state. We propose an image processing method to remove the background region for Baum test. This method uses threshold process, labeling and expansion. It is possible to precisely extract the object region (drawn tree region).
    Download PDF (551K)
  • Tomohiro FUJITA, Toshikazu MATSUI
    Article type: Article
    Session ID: IDY2008-89/HI2008-11
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    It has already been shown that the sharpness evaluation method based on the cooperative human vision model is superior in performance to conventional methods. In order to clarify factors of the vision-model-based method's high performance, the conventional metrics are improved so that they reproduce fundamental characteristics on sharpness in the same way as the vision-model-based metric, and then the performance of the improved metrics is compared with that of the vision-model-based metric. Consequently, the following results were obtained: (1) the improvement was achieved by utilizing the Fourier transform of imaging system's output images instead of the imaging system's MTF; (2) the vision-model-based metric was superior in evaluation performance to the improved metrics. Considering that one large difference between the two kinds of metrics consists in the existence of the image observing mechanism, these results suggest that the model's image observing mechanism can be a factor of the vision-model-based method's high performance.
    Download PDF (881K)
  • Takehito KURIHARA, Naokazu AOKI, Hiroyuki KOBAYASHI
    Article type: Article
    Session ID: IDY2008-90/HI2008-11
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In this research, we investigated the effect of image noise on perceived sharpness. To avoid the difficulty in characterizing natural image, we used as stimuli achromatic 1D sinusoidal grating and 2D uni-frequency pattern, which is one of the simplest forms of image, with various amounts of white Gaussian noise added. Then observers were asked to evaluate their sharpness. For 1D grating stimuli, sharpness seemed independent of noise amount, while, for 2D patterns, it increased with increasing noise. We think this is in line with the observation in our preceding research that sharpness increase by image noise is rather due to enhanced texture than due to enhanced edge.
    Download PDF (451K)
  • Yuzo HISATAKE, Hideki ITO, Masaki OBI, Azusa IKEDA, Yasushi KAWATA, Ak ...
    Article type: Article
    Session ID: IDY2008-91/HI2008-11
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    We have studied the dependency on color chroma and hue for allowable and optimal levels to color difference. The deeper the chroma, the wider the areas of optimal and allowable limits. The plot of each limit for deep RGB on the chromaticity diagram shows that the width for chroma is wider than that of hue area.
    Download PDF (676K)
  • Kazuma SAKAKIBARA, Rei SHIMODA, Tomokazu SHIGA
    Article type: Article
    Session ID: IDY2008-92/HI2008-11
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    With an adaptive dimming technique, the number of gray levels can be increased by dimming the backlight luminance and also by interpolating the original signal. Gray-scale capability of the conventional 8bit LCD is not perfectly favorable to human eye perception. In this report, gray-scale capability of the conventional 8bit LCD was improved by introducing the adaptive dimming technique. It was found that 9-dimming step and 11-bit-equivalent resolution are necessary when the display is used in living room. The experimental results show the effectiveness of the proposed method.
    Download PDF (495K)
  • Article type: Appendix
    Pages App1-
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (78K)
  • Article type: Appendix
    Pages App2-
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (78K)
  • Article type: Appendix
    Pages App3-
    Published: October 09, 2008
    Released on J-STAGE: September 20, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (78K)
feedback
Top