Automatic Estimation of Distribution of Cone Cells to Vessel Locations Using Retina Image

  • JI Yuanting
    Graduate School of Information Science and Electrical Engineering, Kyushu University
  • MOROOKA Ken'ichi
    Graduate School of Information Science and Electrical Engineering, Kyushu University
  • MARTINEZ MOZOS Oscar
    School of Computer Science University of Lincoln
  • TSUJI Tokuo
    Graduate School of Information Science and Electrical Engineering, Kyushu University
  • KURAZUME Ryo
    Graduate School of Information Science and Electrical Engineering, Kyushu University

Bibliographic Information

Other Title
  • 網膜画像を用いた血管位置に対する錐体細胞分布の自動推定(ポスター発表2,計算解剖モデルに基づく診断・治療支援,医用画像処理一般)

Search this article

Description

This paper proposes a method to automatically estimate the distribution patterns of cone cells in a given retina image. As to the spatial relationship between cones and vessels in the retina image, there are three types of the distribution patterns: positive correlation distribution (PCD), negative correlation distribution (NCD), and random distribution (RD). PCD and NCD indicate that the cones tend to be close to or far from the vessels. While the cone cells do not have significant correlation with vessels, the cone distribution is regarded as RD. In our method, a sample image with RD is generated by the vessels extracted from the image, and the virtual cells are selected randomly from the image. Repeating the selection process, many sample images with RDare used to estimate the distribution range of RD. When the distribution of the original cells is above the upper limit or below the lower limit of the RD distribution, the cell distribution is NCD or PCD. Otherwise, the cell distribution is regarded as RD.

Journal

  • IEICE technical report.

    IEICE technical report. 113 (410), 281-284, 2014-01-19

    The Institute of Electronics, Information and Communication Engineers

Details 詳細情報について

  • CRID
    1571698602812238592
  • NII Article ID
    110009821279
  • NII Book ID
    AA11370335
  • Text Lang
    ja
  • Data Source
    • CiNii Articles

Report a problem

Back to top