Leukocytes Segmentation Using Markov Random Fields

  • C. Reta
  • J. A. Gonzalez
  • R. Diaz
  • J. S. Guichard
Part of the Advances in Experimental Medicine and Biology book series (AEMB, volume 696)


The segmentation of leukocytes and their components plays an important role in the extraction of geometric, texture, and morphological characteristics used to diagnose different diseases. This paper presents a novel method to segment leukocytes and their respective nucleus and cytoplasm from microscopic bone marrow leukemia cell images. Our method uses color and texture contextual information of image pixels to extract cellular elements from images, which show heterogeneous color and texture staining and high-cell population. The CIEL  ∗  a  ∗  b  ∗  color space is used to extract color features, whereas a 2D Wold Decomposition model is applied to extract structural and stochastic texture features. The color and texture contextual information is incorporated into an unsupervised binary Markov Random Field segmentation model. Experimental results show the performance of the proposed method on both synthetic and real leukemia cell images. An average accuracy of 95% was achieved in the segmentation of real cell images by comparing those results with manually segmented cell images.


Discrete Fourier Transform Synthetic Image Markov Random Cell Segmentation Harmonic Field 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. 1.
    Kumar B, Joseph D, and Sreenivas T (2002) Teager energy based blood cell segmentation. IEEE 14th Int Conf on Digital Signal Processing 619–622Google Scholar
  2. 2.
    Won C, Nam J, and Choe Y (2004) Segmenting cell images: a deterministic relaxation approach. LNCS 3117:281–291Google Scholar
  3. 3.
    Colantonio S, Gurevich I, and Salvetti O (2007) Automatic fuzzy-neural based segmentation of microscopic cell images. LNCS. doi: 10.1007/978-3-540-76300-0_12Google Scholar
  4. 4.
    Theera-Umpon N (2005) White blood cell segmentation and classification in microscopic bone marrow images. LNCS. doi: 10.1007/11540007_98Google Scholar
  5. 5.
    Dorini L, Minetto R, and Leite N (2007) WBC segmentation using morphological operators and scale-space analysis. IEEE XX Braz Symp on Comput Graphics 294–304Google Scholar
  6. 6.
    Kyungsu K, Jeon J, Choi W et al (2001) Automatic cell classification in human’s peripheral blood images based on morphological image processing. LNCS. doi 10.1007/3-540-45656-2Google Scholar
  7. 7.
    Paschos G (2001) Perceptually uniform color spaces for color texture. IEEE Trans on Image Processing 932–937Google Scholar
  8. 8.
    Francos J (1993) An unified texture model based on a 2-D Wold-Like decomposition. IEEE Trans on Signal Processing 2665–2678Google Scholar
  9. 9.
    Liu F and Picard R (1999). A Spectral 2D Wold Decomposition Algorithm for Homogeneous Random Fields. IEEE Int Conf on Acoustics, Speech and Signal Processing 3501–3504Google Scholar
  10. 10.
    Li S (2003) Modeling Image Analysis Problems Using Markov Random Fields. Stochastic Processes: Modelling and Simulation. North-Holland 473–513Google Scholar
  11. 11.
    Lopez E, Altamirano L (2008) A method based on tree-structured Markov random field and a texture energy function for classification of remote sensing images. IEEE 5th Int Conf on Electrical Engineering, Computing Science and Automatic Control 540–544Google Scholar
  12. 12.
    Morales B, Olmos I, Gonzalez JA, Altamirano L, Alonso J, and Lobato R (2001) Bone marrow smears digitalization. Private medical images collection concluded on December 2001 at Lab de Especialidades del Inst Mexicano del Seguro Social. Puebla, MexicoGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2011

Authors and Affiliations

  • C. Reta
    • 1
  • J. A. Gonzalez
  • R. Diaz
  • J. S. Guichard
  1. 1.National Institute for Astrophysics, Optics, and ElectronicsPueblaMexico

Personalised recommendations