Journal of Printing Science and Technology
Online ISSN : 1882-4935
Print ISSN : 0914-3319
ISSN-L : 0914-3319
Volume 29, Issue 5
Displaying 1-4 of 4 articles from this issue
  • Takeshi AGUI, Kenichi IDE
    1992 Volume 29 Issue 5 Pages 474-479
    Published: September 30, 1992
    Released on J-STAGE: September 27, 2010
    JOURNAL FREE ACCESS
    In fields of commercial designs or simulations of scenes, it is often required to move or remove some indicated objects from pictures. Usually, a layout scanner is used for these operations, in which the operations are executed manually, so it takes much time. To decrease such operations, automatic extraction and paint system is useful. In the present paper, we propose a simple method for automatic extraction of linear objects, and show the experimental results applied to several pictures.
    Download PDF (1496K)
  • Tsuneo TANAKA, Yu KIKU, Saburo YOSHIMURA, Takashi KADOYA
    1992 Volume 29 Issue 5 Pages 480-487
    Published: September 30, 1992
    Released on J-STAGE: September 27, 2010
    JOURNAL FREE ACCESS
    The relationship between acoustic attenuation and liquid penetration into the cellulose nitrate filter was studied as a model system for liquid penetration into paper. Attenuation of ultrasonic wave of 2MHz was measured by means of ultrasonic receiver in process of the liquid penetration. The acoustic attenuation by the membrane with submicrometer pores showed an anomalous behaviour at the initial stage of the penetration for both water and oil. A simplified model for liquid penetration behaviour was applied to give interpretation of the characteristic time dependence of the attenuation level.
    Download PDF (1036K)
  • Shinichi SUZUKI, Takeshi AGUI, Tomoharu NAGAO
    1992 Volume 29 Issue 5 Pages 488-496
    Published: September 30, 1992
    Released on J-STAGE: September 27, 2010
    JOURNAL FREE ACCESS
    Printed matters available on the market are usually warranted by inspection through the human eyes. However, it has demanded to develop an automatic inspection system for labour-savings and standardization of inspection level, because a stable and constant sensuous inspection in a long time cannot be expected. Careless mistakes caused by the inspectors are also unavoidable. The purpose of this study is to develop an automatic inspection system by a neural network for the assurance of printed quality, instead of the human judgement. The neural network system used here is a feed forward and back propagation type. It was proved from the results that the inspection level obtained by the neural network system can approach to that of the human eyes, if the neural network structure concerning a value of the learning parameters and the number of the hidden units are selected appropriately.
    Download PDF (3756K)
  • Yasuhiro ISHINO
    1992 Volume 29 Issue 5 Pages 497
    Published: September 30, 1992
    Released on J-STAGE: September 27, 2010
    JOURNAL FREE ACCESS
    Download PDF (144K)
feedback
Top