An Extraction Method of Cerebral Vessels Based on Multi-Threshold Otsu Classification and Hessian Matrix Enhancement Filtering

Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 269)


An integrated cerebral vascular enhancement method based on the multi-threshold Otsu classification for gray voxels relative to cerebral vessels and the multi-scale Hessian feature for the tubular object enhancement is presented. It implements the multi-threshold Otsu classification to get the cerebral vascular gray voxels, and exploits these voxels’ geometric characteristics by Hessian matrix. And Hessian matrix’s eigenvalues and eigenvectors are used to form a tubular object response function which would be used for further mathematical morphology processing to smooth and mend vessels’ region. Compared with other tubular object enhancement methods, it behaves higher accurateness with stable robustness.


MRA image Cerebral vessels Multi-threshold Otsu classification Hessian matrix Morphology 



This work was supported by National Natural Science Foundation of China (61262031) and Jiangxi Province Graduate Innovation Fund Project (YC2012-S081).


  1. 1.
    Kao Y-H, Teng MMH, Zheng W-Y et al (2010) Removal of CSF pixels on brain MR perfusion images using first several images and Otsu’s thresholding technique. Magn Reson Med 3:743–748Google Scholar
  2. 2.
    Lesage D, Anglini ED, Bloch I et al (2009) A review of 3D vessel lumen segmentation techniques: models, feature and extraction schemes. Med Image Anal 13(6):819–845CrossRefGoogle Scholar
  3. 3.
    Zhang W, Ling H, Prummer S et al (2009) Coronary tree extraction using motion layer separation. Med Image Comput Assist Interv 12(1):116–123Google Scholar
  4. 4.
    Koller TM, Gerig G, Szekely G et al (1995) Multiscale detection of curvilinear structures in 2-D and 3-D image data. In: Proceedings of ICCV vol 2, pp 864–869Google Scholar
  5. 5.
    Hesser J, Männer R, Braus DF et al (1996) Real-time direct volume rendering in functional magnetic resonance imaging. Magn Reson Mater Phys Biol Med 2:87–91Google Scholar
  6. 6.
    Dzyubak OP, Ritman EL (2011) Automation of hessian-based tubularity measure response function in 3D biomedical images. Int J Biomed Imaging 2011:1–16Google Scholar
  7. 7.
    Liu S-P, Chen J (2010) Enhancement method for retinal images based on Gabor filter and morphology. J Optoelectron Laser 21(2):318–322Google Scholar
  8. 8.
    You J, Chen Bo (2011) New approach to retinal image enhancement based on hessian matrix. J Comput Appl 31(6):1560–1562Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2014

Authors and Affiliations

  1. 1.School of Basic ScienceEast China Jiaotong UniversityNanchangChina

Personalised recommendations