2次元フーリエ変換による肝表面像の解析 I 基礎的検討

DOI

書誌事項

タイトル別名
  • Analysis of liver surface structure using 2-dimensional Fourier transform.
  • ―(I)基礎的検討―

抄録

For the diagnosis of chronic liver disease, peritoneoscopic findings of liver surface image are as important as histopathological findings of liver biopsy, especially when the lesions become coarse and inhomogeneous with the progress of the disease. The peritoneoscopic findings, however, tend to depend at least somewhat upon operator's experience, and it is sometimes difficult to obtain objective and quantitative findings.<BR> In the present paper, we studied computer image processing of 2-dimensional Fourier transform, in order to get more objective and quantitative findings of liver surface structure taken by electroperitoneoscopy. After the shading correction of peritoneoscopic image, image data was converted to power spectrum by 2-dimensional Fourier transform. Converting the resulting data to polar coordinate system, spatial frequency was extracted and the frequential distribution of liver surface particle was calculated. Then, the scale calibration was carried out, and the distribution of particle length was obtained.<BR> For the basal approach to check the analyzing procedure, we first used the liver surface particle models, one type of which was made of red sphere particles and was photographed by peritoneoscope, and another type of which was drawn by computer simulation. Secondly, we analyzed clinical liver surface image taken by peritoneoscopy.<BR> The results of model examination indicated that the particle frequency and the particle length obtained by this analyzing method had a good correspondence with the practical value of the particle in each model. In the analysis of clinical image of liver cirrhosis, particle element reflecting the size and distribution of liver surface nodule was detected. We considered that this method was useful to obtain objective and quantitative findings of liver surface image.

収録刊行物

詳細情報 詳細情報について

  • CRID
    1390282679191096448
  • NII論文ID
    130004252885
  • DOI
    10.11280/gee1973b.32.2342
  • ISSN
    18845738
    03871207
  • 本文言語コード
    ja
  • データソース種別
    • JaLC
    • CiNii Articles
  • 抄録ライセンスフラグ
    使用不可

問題の指摘

ページトップへ