Skip to main content
SpringerLink
Log in

Evaluation and analysis of data driven in expectation maximization segmentation through various initialization techniques in medical images

  • Published:
Multimedia Tools and Applications Aims and scope Submit manuscript

Abstract

The operation of partitioning an image into a collection of connected sets of pixels is known as image segmentation. Expectation Maximization (EM) segmentation requires appropriate initialization of tissue class mean and variance, as it is get stuck on particular possible area of intensity of the probability nature. Decisive initialization is a necessary exploratory process for the forthcoming convergence of the algorithm to best regional maximum of possibility task. Indiscriminate initialization is not dossier directed, deep from optimal, outcomes are not reproducible, do not take advantage of deep rooted patterns in the data or may be loaded on outliers. This paper evaluates the performance of EM segmentation with random initialization, histogram guided initialization and initialization with k-means with respect to computational complexity and root mean squared error of tissue class mean and variance, updated by EM, with manually estimated tissue class mean and variance as ground truth, on paramount plane, T1 contrast and Magnetic resonance (MR) images of Glioblastoma Multiframe (GBM) Edema complex. The random initialization and histogram guided initialization was experimented for k-means, from the clustered output of which; initial tissue class mean and variance for EM are derived. RMS error remains the same for EM initialized and histogram guided K-means. EM initialized with K-means which has histogram guided initialization converges fast than the random initialization K-means, but the computational time is more for the former initialization than the latter. The experimental evaluation of EM initialization schemes and Fuzzy Cluster Means (FCM) were performed in MATLAB. The efficacy of Fuzzy Cluster Means clustering was analyzed qualitatively on tumour-edema complex. FCM could identify only 3 classes including background in the MR specimens. FCM consider edema and certain parts of WM as a single tissue class. Similarly, FCM clubs GM, CSF and necrotic focus into tissue class and produced empty clusters.

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

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Similar content being viewed by others

References

  1. Dempster AP, Laird NM, Rubin DB (1977) Maximum likelihood estimation from incomplete data via the EM algorithm. J R Stat Soc 39(1):1–38

    MATH  Google Scholar 

  2. Fatakdawala H, Xu J, Basavanhally A, Bhanot G, Ganesan S, Feldman M, Tomaszewski JE, Madabhushi A (2010) Expectation– Maximization-Driven Geodesic Active Contour with Overlap Resolution (EMaGACOR): application to lymphocyte segmentation on breast cancer histopathology. IEEE Trans Biomed Eng 57(7):1676–1689. https://doi.org/10.1109/TBME.2010.2041232

    Article  Google Scholar 

  3. Fwu JK, Djuric PM (1997) EM algorithm for image segmentation initialized by a tree structure scheme. IEEE Trans Image Process 6(2):349–352

    Article  Google Scholar 

  4. Gonzalez R, Woods R (1992) Digital image processing. Addison-Wesley Publishing Company, Boston, p 191

    Google Scholar 

  5. Gooya A, Pohl KM, Bilello M, Cirillo L, Biros G, Melhem ER, Davatzikos C (2012) GLISTR: Glioma Image Segmentation and Registration. IEEE Trans Med Imaging 31(10):1941–1954. https://doi.org/10.1109/TMI.2012.2210558

    Article  Google Scholar 

  6. Greensp H, Ruf A, Goldberger J (2006) Constrained Gaussian mixture model framework for automatic segmentation of MR brain images. IEEE Trans Med Imaging 25(9):1233–1245

    Article  Google Scholar 

  7. Ilea DE, Whelan PF (2006) Color image segmentation using a self- initializing EM algorithm. Proc. 6th international conference on visualization, imaging and image processing, 28–30 August 2006, Palma De Mallorca, ISBN 0–88986–598-1 Source: OAI

  8. Karthikeyan A, Kumar PS (2017) Cluster comput. https://doi.org/10.1007/s10586-017-0979-0

  9. Karthikeyan A, Senthil Kumar P (2017) Randomly prioritized buffer-less routing architecture for 3D network on chip. Comput Electr Eng 59:39–50. https://doi.org/10.1016/j.compeleceng.2017.03.006 ISSN 0045-7906

    Article  Google Scholar 

  10. Khan S, Ahmad A (2004) Cluster center initialization algorithm for K-means clustering. Pattern Recogn Lett 25(11):1293–1302

    Article  Google Scholar 

  11. Liang Z, Wang S (2009) An EM approach to MAP solution of segmenting tissue mixtures: a numerical analysis. IEEE Trans Med Imaging 28(2):297–310

    Article  MathSciNet  Google Scholar 

  12. Lomer JB, Bujna K (2013) Simple methods for initializing the EM algorithm for Gaussian mixture models. Computing Research Repository, http://arxiv.org/abs/1312.5946

  13. Lu C-F, Wang P-S, Chou Y-C, Li H-C, Soong B-W, Wu Y-T (2008) Segmentation of diffusion-weighted brain images using expectation maximization algorithm initialized by hierarchical clustering. Engineering in Medicine and Biology Society, EMBS 2008. 30th Annual International Conference of the IEEE, pp. 5502–5505, 20–25 Aug. 2008, https://doi.org/10.1109/IEMBS.2008.4650460

  14. Lynch M, Ilea D, Robinson K, Ghita O, Whelan PF (2007) Automatic seed initialization for the expectation-maximization algorithm and its application in 3D medical imaging. J Med Eng Technol 31(5):332–340

    Article  Google Scholar 

  15. Melnykova V, Melnykovb I (2012) Initializing the EM algorithm in Gaussian mixture models with an unknown number of components. Comput Stat Data Anal 56(6):1381–1395

    Article  MathSciNet  Google Scholar 

  16. Mitra P, Pal SK, Siddiqi MA (2003) Non-convex clustering using expectation maximization algorithm with rough set initialization. Pattern Recogn Lett 24(6):863–873

    Article  MATH  Google Scholar 

  17. Otsu N (1979) A threshold selection method from gray-level histograms. IEEE Trans Syst Man Cybern 9(1):62–66

    Article  MathSciNet  Google Scholar 

  18. Pal SK, Mitra P (2002) Multispectral image segmentation using the rough- set-initialized EM algorithm. IEEE Trans Geosci Remote Sens 40(11):2495–2501

    Article  Google Scholar 

  19. Ray S, Turi RH (1999) Determination of number of clusters in-means clustering and application in colour image segmentation. Proc 4th International Conference on Advances in Pattern Recognition and Digital Techniques (ICAPRDT'99), Calcutta, pp. 137–143

  20. Shen S, Sandham W, Granat M, Sterr A (2005) MRI fuzzy segmentation of brain tissue using neighborhood attraction with neural-network optimization. IEEE Trans Inf Technol Biomed 9(3):459–467. https://doi.org/10.1109/TITB.2005.847500

    Article  Google Scholar 

  21. Srinivasan P, Shenton ME, Bouix S, EM segmentation: automatic tissue class intensity initialization using K-means. http://hdl.handle.net/10380/3300

  22. Tian GJ, Xia Y, Zhang Y, Feng D (2011) Hybrid genetic and variational expectation-maximization algorithm for gaussian-mixture-model- based brain MR image segmentation. IEEE Trans Inf Technol Biomed 15(3):373–380. https://doi.org/10.1109/TITB.2011.2106135

    Article  Google Scholar 

  23. Tomasi C, Manduchi R (1998) Bilateral filtering for gray and color images. Proceedings of the 1998 I.E. international conference on computer vision, Bombay

  24. Vlassis N, Likas A (2002) A greedy EM algorithm for Gaussian mixture learning. Neural Process Lett 1(1):77

    Article  MATH  Google Scholar 

  25. Volkovich Z, Avros R., Golani M (2011) On initialization of the expectation maximization clustering algorithm. Transaction on Evolutionary Algorithm and Clustering Online Publication, June 2011, www.pcoglobal.com/gjto.htm, CA-O34/GJTO, ISSN: 2229–8711

  26. Wang S, Lu H, Liang Z (2009) Theoretical solution to MAP-EM partial volume segmentation of medical images. Int J Imaging Syst Technol 19(2):111–119

    Article  Google Scholar 

  27. Wu YT, Chou YC, Lu CF, Huang SR, Guo WY (2009) Tissue classification from brain perfusion MR images using expectation- maximization algorithm initialized by hierarchical clustering on whitened data. 11th international conference on biomedical engineering, IFMBE proceedings vol. 23, pp 714–717

  28. Zuiderveld K (1994) Contrast limited adaptive histogram equalization. Graphic Gems IV. San Diego: Academic Press Professional, pp 474–485

Download references

Author information

Authors and Affiliations

  1. Department of Electronics and Communication Engineering, Vel Tech Multi Tech Dr. Rangarajan Dr. Sakunthala Engineering College, Chennai, India

    V. Prabhu, P. G. Kuppusamy & A. Karthikeyan

  2. Department of Electronics and Communication Engineering, Sri Ramanujar Engineering College, Chennai, India

    R. Varatharajan

Authors
  1. V. Prabhu
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. P. G. Kuppusamy
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. A. Karthikeyan
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. R. Varatharajan
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to V. Prabhu.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Prabhu, V., Kuppusamy, P.G., Karthikeyan, A. et al. Evaluation and analysis of data driven in expectation maximization segmentation through various initialization techniques in medical images. Multimed Tools Appl 77, 10375–10390 (2018). https://doi.org/10.1007/s11042-018-5792-0

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11042-018-5792-0

Keywords

Search

Navigation