Skip to main content
SpringerLink
Log in

Fast medical image mixture density clustering segmentation using stratification sampling and kernel density estimation

  • Original Paper
  • Published:
Signal, Image and Video Processing Aims and scope Submit manuscript

Abstract

The Gaussian mixture models (GMMs) is a flexible and powerful density clustering tool. However, the application of it to medical image segmentation faces some difficulties. First, estimation of the number of components is still an open question. Second, the speed of it for large medical image is slow. Moreover, GMMs has the problem of noise sensitivity. In this paper, the kernel density estimation method is used to estimate the number of components K, and three strategies are proposed to improve the segmentation speed of GMMs. First, a histogram stratification sampling strategy is proposed to reduce the size of the training data. Second, a binning strategy is proposed to search the neighbor points of each center data to compute the approximate density function of the samples. Third, a hill-climbing algorithm with the dynamic step size is designed to find the local maxima of the density function. The kernel density estimation method and sampling technology reduce the effect of noise. Experimental results with the simulated brain images and real CT images show that the proposed algorithm has better performance in generating explainable segmentations with faster speed than the common GMMs algorithm.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Abbreviations

X :

A medical image

{x 1, x 2, . . . x N }:

All pixels of X

N :

The number of pixels

N h :

The size of the hth stratum of X

n h :

Size of the hth stratum of SX

W h :

Proportion of the hth stratum of X

w h :

Proportion of the hth stratum of SX

N :

The Size of sample SX

K :

The number of components of mixture models

α r :

Is the weight of component r

μ r :

Is the mean of component r

\({\sum _r}\) :

The covariance of component r

\({\varphi _r =\{\alpha _r ,\mu _r ,\Sigma _r \}}\) :

All parameters of component r

\({\theta=(\alpha_1,\alpha_2,\ldots,\,\alpha_K,\mu_1,\mu_2,\ldots\mu_K,\,\sum _1 ,\sum _2 ,\ldots\sum _K ) }\) :

All parameters of mixture model

\({f_r (x|\mu _r ,\sum _r )}\) :

The density function of component r

log L(x|θ):

Log-likelihood function

SX :

The samples extracted from a medical image X

SY = {Y1 , . . . , Y n }:

2D data set converted from the 3D data set SX

G = {Lmin , Lmin + 1, . . . , Lmax}:

The gray level set of a medical image X

p :

The number of sampling strata

His:

The histogram of image X

SHis :

The smoothed histogram

\({\overline{Y}}\) :

The mean of image X

\({\overline{y}_h}\) :

The mean of the hth stratum of sample SX

α :

The precision

δ i :

The ith step optimum size of hill-climbing procedure

\({\tilde {f}(Y_i )}\) :

Approximate density

\({\nabla \tilde {f}(Y_i )}\) :

The gradient function

\({\nabla ^{2} \tilde {f}(Y_i )}\) :

Hessian matrix of density function

S :

Unit gradient vector

\({N_{p\cap g} (r)}\) :

The number of pixels classified by both the proposed method and the ground truth as model r

N p (r):

The number of pixels classified as model r by the proposed method

N g (r):

The number of pixels classified as model r by the ground truth

ξ :

Thresholding

σ :

Window width

xw :

The bin width of x-coordinate

yw :

The bin width of y-coordinate

Y :

The data in the same bin or neighbor bins of Y i

s :

The size of Y

References

  1. Zhang D.Q., Chen S.C.: A novel kernelized fuzzy C-means algorithm with application in medical image segmentation. Artif. Intell. Med. 32(1), 37–50 (2004)

    Article  Google Scholar 

  2. Bricq S., Collet C.H., Armspach J.P.: Unifying framework for multimodal brain MRI segmentation based on Hidden Markov chains. Med. Image. Anal. 12(6), 639–652 (2008)

    Article  Google Scholar 

  3. Pham D.L., Xu C.: Current methods in medical image segmentation. Ann. Biomed. Eng. 2(8), 315–338 (2000)

    Google Scholar 

  4. Zhang H., Jason E.F., Sally A.G.: Image segmentation evaluation: a survey of unsupervised methods. Comput. Vis. Image. Und. 110(2), 260–280 (2008)

    Article  Google Scholar 

  5. Marroquin J.L., Vemuri B.C., Botello S., Calderon F.: An accurate and efficient Bayesian method for automatic segmentation of brain MRI. IEEE Trans. Med. Imag. 21(8), 934–945 (2002)

    Article  Google Scholar 

  6. Chao W.H., Chen Y.Y., Lin S.H.: Automatic segmentation of magnetic resonance images using a decision tree with spatial information. Comput. Med. Imag. Grap. 33(2), 111–121 (2009)

    Article  Google Scholar 

  7. Iscan Z., Yüksel A., Dokur Z.: Medical image segmentation with transform and moment based features and incremental supervised neural network. Digit. Signal Process 19(5), 890–901 (2009)

    Article  Google Scholar 

  8. Tang H., Dillenseger J.L., Bao X.D., Luo L.M.: A vectorial image soft segmentation method based on neighborhood weighted Gaussian mixture model. Comput. Med. Imag. Grap. 33(8), 644–650 (2009)

    Article  Google Scholar 

  9. Jain A.K., Murty M.N., Flynn P.J.: Data clustering: a review. ACM Comput. Surveys 31(3), 264–323 (1999)

    Article  Google Scholar 

  10. Kuo W.F., Lin C.Y., Sun Y.N.: Brain MR images segmentation using statistical ratio: Mapping between watershed and competitive Hopfield clustering network algorithms. Comput. Meth. Prog. Bio. 91(3), 191–198 (2008)

    Article  Google Scholar 

  11. Bezdek J.C.: Pattern recognition with fuzzy objective function algorithms. SIAM Rev. 25(3), 442–442 (1983)

    Google Scholar 

  12. Cai W.L., Chen S.C., Zhang D.Q.: Fast and robust fuzzy c-means clustering algorithms incorporating local information for image segmentation. Pattern Recogn. 40(3), 825–838 (2007)

    Article  MATH  Google Scholar 

  13. Yong Y.: Image segmentation based on fuzzy clustering with neighborhood information. Opt. Appl. 39(1), 135–147 (2009)

    Google Scholar 

  14. Guillemaud R., Brady M.: Estimating the bias field of MR images. IEEE Trans. Med. Imaging 16(3), 238–251 (1997)

    Article  Google Scholar 

  15. Wells W.M., Grimso W.E.L., Kikinis R.: Adaptive segmentation of MRI data. IEEE Trans. Med. Imaging 15(4), 429–442 (1996)

    Article  Google Scholar 

  16. Hayit G., Adi T.P.: Medical image categorization and retrieval for PACS using the GMM-KL framework. IEEE T. Inf. Technol. B 11(2), 190–202 (2007)

    Article  Google Scholar 

  17. Dempster A.P., Laird N.M., Rubin D.B.: Maximum-likelihood from incomplete data via the EM algorithm. J. R. Stat. Soc. Ser. B 39(1), 1–38 (1977)

    MathSciNet  MATH  Google Scholar 

  18. Tang Y.G., Liu D., Guan X.P.: Multi-resolution image segmentation based on Gaussian mixture model. J. Syst. Eng. Electron 17(4), 870–874 (2006)

    Article  MATH  Google Scholar 

  19. Yang X.Y., Shankar M.: Image segmentation using finite mixtures and spatial information. Image Vision Comput. 22(9), 735–745 (2004)

    Article  Google Scholar 

  20. Khayati R., Vafadust M., Towhidkhah F.: Fully automatic segmentation of multiple sclerosis lesions in brain MRFLAIR images using adaptive mixtures method and Markov random field model. Comput. Biol. Med. 38(3), 379–390 (2008)

    Google Scholar 

  21. Adelino R.F.D.S.: A Dirichlet process mixture model for brain MRI tissue classification. Med. Image Anal. 11(2), 169–182 (2007)

    Article  Google Scholar 

  22. McLachlan G., Peel D.: Finite Mixture Models. Wiley series in probability and statistics, Chap. 1, pp. 1–20. Wiley, New York (2000)

    Google Scholar 

  23. Parzen E.: On estimation of a probability density function and mode. Ann. Math. Stat. 35(3), 1065–1076 (1962)

    Article  MathSciNet  Google Scholar 

  24. Govindarajulu, Z.: Elements of Sampling Theory and Methods, Pearson Education, Chap. 5. pp. 75–105 (1999)

  25. Laurens M.V.D., Geoffrey H.: Visualizing data using t-SNE. J. Mach. Learn. Res. 9(1), 2579–2605 (2008)

    Google Scholar 

  26. Figueiredo M.A.T., Jain A.K.: Unsupervised learning of finite mixture models. IEEE T. Pattern Anal. 24, 381–396 (2002)

    Article  Google Scholar 

  27. http://www.bic.mni.mcgill.ca/brainweb

  28. Zijdenbos A.P., Dawant B.M., Margolin R.A., Palmer A.C.: Morphometric analysis of white matter lesions in MR images: method and validation. IEEE Trans. Med. Imaging 13(4), 716–724 (1994)

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. School of Computer Science and Telecommunication, Jiangsu University, 212013, Zhenjiang, Jiangsu Province, Peoples’ Republic of China

    Cong-Hua Xie, Yu-Qing Song & Jian-Mei Chen

  2. School of Computer Science and Engineering, Changshu Institute of Technology, 215500, SuZhou, Jiangsu Province, Peoples’ Republic of China

    Cong-Hua Xie

Authors
  1. Cong-Hua Xie
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Yu-Qing Song
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Jian-Mei Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Cong-Hua Xie.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Xie, CH., Song, YQ. & Chen, JM. Fast medical image mixture density clustering segmentation using stratification sampling and kernel density estimation. SIViP 5, 257–267 (2011). https://doi.org/10.1007/s11760-010-0159-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11760-010-0159-7

Keywords

Search

Navigation