Formation of feature spaces based on tortuosity estimation of vessel central lines in the analysis problem of abnormal changes in the structure of the fundus vascular system
A.O. Korepanov, M.A. Ananyin

Image Processing Systems Institute оf the RAS,
Samara State Aerospace University (SSAU)

Full text of article: Russian language.

Abstract:
We develop a method of formation of feature spaces with regard to the problem of classification of the fundus blood vessels. The proposed method is based on tortuosity estimation of pre-treated vessel central lines with the use of smoothing filters with different bandwidths (with various degrees of smoothing) that results in the use of the wavelet transform when calculating differential characteristics. We analyze the efficiency of proposed features and compare them with existing methods of formation of feature spaces.

Key words:
fundus vascular system, vessel central lines, feature space formation, classification.

Citation:
Korepanov AO, Ananyin MA. Formation of feature spaces based on tortuosity estimation of vessel central lines in the analysis problem of abnormal changes in the structure of the fundus vascular system [In Russian]. Computer Optics 2007; 31(1): 52-57.

References:

  1. Ilyasova NYu, Ustinov AV, Baranov VG. An expert computer system for diagnosing eye diseases from retina images [In Russian]. Computer Optics 1999; 19: 202-209.
  2. Hoover A, Kouznetsova V, Goldbaum M. Locating blood vessels in retinal images by piece-wise threshold probing of a matched filter response. IEEE Transactions on Medical Imaging 2000: 19(3): 203-210.
  3. Martinez-Perez ME, Hughes AD, Stanton AV, Thom SA, Bharath AA, Parker KH. Retinal Blood Vessel Segmentation by means of Scale-Space Analysis and Region Growing. Proceedings of the Second International Conference on Medical Image Computing and Computer Assisted Intervention. Cambridge, 1999.
  4. Hart WE. Measurement and classification of retinal vascular tortuosity. Applied Numerical Mathematics Dept, Sandia National Laboratories.
  5. Grisan Enrico, Foracchia Marco, Ruggeri Alfredo. A novel method for the automatic evaluation of retinal vessel tortuosity. Department of Information Engineering, University of Padova, Italy.
  6. Kylstra JA, Freedman SF, Capowsky JJ, Hall JG. Measurement of retinal vessel tortuosity as a means of diagnosis plus disease. ROP Invest Ophthalmol Vis Sci 1995; 36: 77
  7. Dubrovin BA, Novikov SP, Fomenko AT. Modern geometry: methods and applications [In Russian]. Moscow: Editorial URSS, 1998; 1.
  8. Ayvazyan SA (ed.). Applied statistics: classification and dimension reduction [In Russian]. Moscow: Finance and Statistics Publisher, 1987.
  9. Ilyasova NYu, Korepanov AO, Chikulaev PM. A method for extracting blood vessel central lines in diagnostic images [In Russian]. Computer Optics 2006; 29: 146-151.
  10. Fukunaga K. Introduction to statistical pattern recognition [In Russian]. Moscow: “Nauka” Publisher, 1979.

© 2009, ИСОИ РАН
Россия, 443001, Самара, ул. Молодогвардейская, 151; электронная почта: ko@smr.ru ; тел: +7 (846) 332-56-22, факс: +7 (846 2) 332-56-20