Reports of the Technical Conference of the Institute of Image Electronics Engineers of Japan
Online ISSN : 2758-9218
Print ISSN : 0285-3957
Reports of the 234th Technical Conference of the Institute of Image Electronics Engineers of Japan
Displaying 1-6 of 6 articles from this issue
Date:November 2, 2007 Location:Osaka University
11:00-12:00 Chair: Taro IMAGAWA, Matsushita Electric Industrial Co., Ltd.
  • Masaru KUBOTA, Katsuyuki SHINOHARA
    Session ID: 07-02-01
    Published: 2007
    Released on J-STAGE: March 11, 2008
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    This paper proposed the image retrieval system that uses the sketch retrieval and SOM. Currently, a lot of image retrieval system might need a similar image to retrieve. Moreover, user was easy to fall into the local solution because image retrieval system was only to display retrieval result by the ranking form. Then, this paper obtained high retrieval efficiency by mounting the sketch retrieval and SOM. The sketch retrieval mounted to retrieve by the vague imagination in user's head. Moreover, SOM mounted to browsing in the database. In addition, this system was a quick retrieval because the increment retrieval mounted.
    Download PDF (607K)
  • Hirokazu TOMONAGA, Takuya ABE, Katsyuki SHINOHARA
    Session ID: 07-02-02
    Published: 2007
    Released on J-STAGE: March 11, 2008
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    In this report, we attempted to extract features that are effective to retrieval the photo image to construct CBIR for website of sharing or stocking photos. Concretely, we defined 46 region segmentation patterns from the positions presumed as main subjects of basic compositions of the photograph. Next, we choose the region having degree of visual attention high. And we used the similarity that weights high this region at retrieval. We compared the precision of our proposing method with Color layout. This experiment showed that the precision of our proposing method is high than Color layout, and we found that our proposing method was effective.
    Download PDF (442K)
13:30-14:00 Chair: Masutaka INOUE, Sanyo Electric Co., Ltd.
  • Mikiya HIRONAGA, Noriyuki SHIMANO
    Session ID: 07-02-03
    Published: 2007
    Released on J-STAGE: March 11, 2008
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    It is very important to recover spectral reflectances of imaged objects accurately for the record of the precious art paintings in digital archive and for the reproduction of them under variety illuminants. The accuracy of the recovery depends on the recovery models and on the spectral sensitivities of sensors. Although the Wiener estimate gives very accurate recovery of the reflectances, prior knowledge of spectral reflectances about the several points on the imaged painting is required to estimate the noise variance of an image acquisition system. We have already proposed a new method to estimate the noise variance of the system and to recover spectral reflectances of the art paintings by using the estimated variance. In this paper, we report the experimental results of the color reproduction of many art paintings of which spectral reflectances are recovered by the proposal.
    Download PDF (376K)
14:10-15:10 Chair: Hidekazu SEKIZAWA, Toshiba Tec Document Prosessing Systems Co.,Ltd.
15:20-16:00 Chair: Taro IMAGAWA, Matsushita Electric Industrial Co., Ltd.
  • Yoshifumi KITAMURA
    Session ID: 07-02-06
    Published: 2007
    Released on J-STAGE: March 11, 2008
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Two unique digital tables which effectively promote cooperative interactions for co-located multiple users are introduced. The IllusionHole is an interactive display system that allows three or more moving observers to simultaneously observe stereoscopic image pairs from their own viewpoints. With a simple configuration, it provides intelligible 3D stereoscopic images free of flicker and distortion. The SharedWell is a system designed to improve strategic negotiation by supporting seamless and interactive control of public and private information spaces by naturally changing their relative spatial location.
    Download PDF (517K)
feedback
Top