This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
A Hybrid Harmony Search Algorithm to MRI BrainSegmentation
Abstract—Automatic brain MRI image segmentation is achallenging problem and received significant attention in the fieldof medical image processing. In this paper, we present a newdynamic clustering algorithm based on the Harmony Search (HS)hybridized with Fuzzy C-means called DCHS to automaticallysegment the brain MRI image in an intelligent manner. Inthis algorithm, the capability of standard HS is modified toautomatically evolve the appropriate number of clusters as wellas the locations of cluster centers. By incorporating the conceptof variable length in each harmony memory vector, DCHS isable to encode variable numbers of candidate cluster centersat each iteration. Furthermore, a new HS operator, called the‘empty operator’ is introduced to support the selection of emptydecision variables in the harmony memory vector. The PBMFcluster validity index is used as an objective function to validatethe clustering result obtained from each harmony memory vector.The proposed algorithm is applied on several simulated T1-weighted normal and MS lesion magnetic resonance brain images.The experimental results show the ability of DCHS to find the ap-propriate number of naturally occurring regions in brain images.Furthermore, superiority of the proposed algorithm over differentclustering-based algorithms is demonstrated quantitatively. Allthe segmented results obtained by DCHS are also compared withthe available ground truth images.
Index Terms—Automatic Brain MRI segmentation, dynamicfuzzy clustering, harmony search, PBMF index
I. INTRODUCTION
The segmentation of Magnetic Resonance Imaging (MRI)
brain images is a fundamental process for medical research and
clinical applications such as quantification of tissue volume,
diagnose illnesses, aid in computer-guided surgery, treatment
Fig. 6. (a) original multiple sclerosis lesions T1- brain image in z1 plane (b)segmented result obtained by DCHS (6 clusters).
In order to compare the performance of the proposed algo-
rithm with other dynamic clustering-based algorithms, Table
II shows the results of the same experiment conducted by
FVGAPS [42], [43] and FVGA [45], [46], where these results
reported in [42], [43]. FVGA algorithm also is a dynamic
clustering algorithm that can automatically determine number
of clusters with the appropriate cluster center values. FVGA
is based on Genetic Algorithm combined with XB index as
an objective function. Table II shows the number of clusters
that FVGAPS found and the corresponding MS rates. Also
the number of clusters found by FVGA for the given images
and their corresponding MS rates are shown in the same table.
By looking at these findings, the number of clusters found
by DCHS is much better and more accurate than these found
by FVGAPS and FVGA. FVGAPS unable to find the optimal
number of clusters for all given images except image z10 only.
While FVGA failed to find the optimal number of clusters
for all given images. The MS rates also show the superiority
of the proposed DCHS algorithm compared with the other
algorithms.
V. CONCLUSION
We have presented in this paper a novel dynamic clustering
algorithm called DCHS based on Harmony Search algorithm
hybridized with Fuzzy C-means algorithm. DCHS has the
ability to cluster the given data set automatically without any
prior knowledge of number of clusters that may the given
data set has. DCHS have been used as an image segmentation
algorithm to dynamically segment a simulated normal and
multiple sclerosis lesions brain MRI images.
The numbers of clusters found by DCHS for the brain im-
ages were the best among other dynamic clustering algorithms
(FVGAPS and FVGA). The accuracy rates of segmentation
results were the highest for multiple sclerosis lesions images
while comparable in the normal brain images.
Regarding the MRI artifacts such as noise and intensity non-
uniformity, the upcoming research will focus on how to im-
prove the segmentation results of DCHS algorithm and how to
incorporate the spatial information to reduce the effectiveness
of these artifacts. Furthermore, another study of how to reduce
the impact of the problem of non-similar brain tissue types
(a) (b)
Fig. 7. (a) original multiple sclerosis lesions T1- brain image in z5 plane (b)segmented result obtained by DCHS (6 clusters).
with similar intensity appearance will be also the upcoming
research interest.
ACKNOWLEDGMENT
The authors thank Dr. Dhanesh Ramachandram and Dr. Sri-
parna Saha for their comments regarding this manuscript. This
research is supported by Universiti Sains Malaysia, USM’s
fellowship scheme and ’Universiti Sains Malaysia Research
University Grant’ grant titled ’Delineation and visualization of
Tumour and Risk Structures - DVTRS’ under grant number
1001/PKOMP/817001.
REFERENCES
[1] W. Dou, S. Ruan, Y. Chen, D. Bloyet, and J.-M. Constans, “A frameworkof fuzzy information fusion for the segmentation of brain tumor tissueson mr images,” Image and Vision Computing, vol. 25, no. 2, pp. 164–171,2007.
[2] J. Wang, J. Kong, Y. Lu, M. Qi, and B. Zhang, “A modified fcmalgorithm for mri brain image segmentation using both local and non-local spatial constraints,” Computerized Medical Imaging and Graphics,vol. 32, no. 8, pp. 685–698, 2008.
[3] A. W.-C. Liew and H. Yan, “Current methods in the automatic tissuesegmentation of 3d magnetic resonance brain images,” Current medicalimaging reviews, vol. 2, pp. 91–103, 2006.
[4] I. Wells, W. M., W. E. L. Grimson, R. Kikinis, and F. A. Jolesz, “Adaptivesegmentation of mri data,” Medical Imaging, IEEE Transactions on,vol. 15, no. 4, pp. 429–442, 1996.
[5] T. Kapur, W. E. L. Grimson, W. M. Wells, and R. Kikinis, “Segmen-tation of brain tissue from magnetic resonance images,” Medical ImageAnalysis, vol. 1, no. 2, pp. 109–127, 1996.
[6] J. Zhou and J. C. Rajapakse, “Fuzzy approach to incorporate hemo-dynamic variability and contextual information for detection of brainactivation,” Neurocomputing, vol. 71, no. 16-18, pp. 3184–3192, 2008.
[7] L. Szilagyi, Z. Benyo, S. M. Szilagyi, and H. S. Adam, “Mr brainimage segmentation using an enhanced fuzzy c-means algorithm,” inProceedings of the 25th Annual International Conference of the IEEEEngineering in Medicine and Biology Society,, vol. 1, 2003, pp. 724–726.
[8] H. A. Mokbel, M. E.-S. Morsy, and F. E. Z. Abou-Chadi, “Automaticsegmentation and labeling of human brain tissue from mr images,” inSeventeenth National Radio Science Conference, 17th NRSC ’, 2000, pp.1–8.
[9] L. Xiaohe, Z. Taiyi, and Q. Zhan, “Image segmentation using fuzzyclustering with spatial constraints based on markov random field viabayesian theory,” IEICE Trans. Fundam. Electron. Commun. Comput.Sci., vol. E91-A, no. 3, pp. 723–729, 2008.
[10] K. Van Leemput, F. Maes, D. Vandermeulen, and P. Suetens, “Automatedmodel-based tissue classification of mr images of the brain,” IEEETransactions on Medical Imaging, vol. 18, no. 10, pp. 897–908, 1999.
[11] K. Van Leemput, F. Maes, D. Vandermeulen, and P. Suetens, “Automatedmodel-based bias field correction of mr images of the brain,” IEEETransactions on Medical Imaging, vol. 18, no. 10, pp. 885–896, 1999.
[12] J. C. Bezdek, L. O. Hall, and L. P. Clarke, “Review of mr imagesegmentation techniques using pattern recognition,” Medical Physics,vol. 20, no. 4, pp. 1033–1048, 1993.
[13] Y. L. Chang and X. Li, “Adaptive image region-growing,” IEEE Trans-actions on Image Processing, vol. 3, no. 6, pp. 868–872, 1994.
[14] R. Adams and L. Bischof, “Seeded region growing,” IEEE Transactionson pattern analysis and machine intelligence, vol. 16, no. 6, pp. 641–647,1994.
[15] R. Pohle and K. D. Toennies, “Segmentation of medical images usingadaptive region growing,” in Proceedings of SPIE (Medical Imaging),vol. 4322, San Diego, 2001, pp. 1337–1346.
[16] J. Sijbers, P. Scheunders, M. Verhoye, A. Van der Linden, D. van Dyck,and E. Raman, “Watershed-based segmentation of 3d mr data for volumequantization,” Magnetic Resonance Imaging, vol. 15, no. 6, pp. 679–688,1997.
[17] M. S. Atkins and B. T. Mackiewich, “Fully automatic segmentation ofthe brain in mri,” IEEE Transactions on Medical Imaging,, vol. 17, no. 1,pp. 98–107, 1998.
[18] M. Ashtari, J. L. Zito, B. I. Gold, J. A. Lieberman, M. T. Borenstein, andP. G. Herman, “Computerized volume measurement of brain structure,”Investigative Radiology, vol. 25, no. 7, pp. 798–805, 1990.
[19] L. Ji and H. Yan, “Attractable snakes based on the greedy algorithm forcontour extraction,” Pattern Recognition, vol. 35, no. 4, pp. 791–806,2002.
[20] T. McInerney and D. Terzopoulos, “Deformable models in medical imageanalysis: a survey,” Medical Image Analysis, vol. 1, no. 2, pp. 91–108,1996.
[21] Y. Zhou and J. Bai, “Atlas-based fuzzy connectedness segmentationand intensity nonuniformity correction applied to brain mri,” IEEETransactions on Biomedical Engineering,, vol. 54, no. 1, pp. 122–129,2007.
[22] M. C. Clark, “Knowledge-guided processing of magnetic resonanceimages of the brain,” Ph.D. dissertation, University of South Florida,Department of Computer Science and Engineering, South Florida USA,1998.
[23] M. Clark, L. Hall, D. Goldgof, and M. Silbiger, “Using fuzzy informationin knowledge guided segmentation of brain tumors,” in Fuzzy Logic inArtificial Intelligence Towards Intelligent Systems, 1997, pp. 167–181.
[24] M. Sonka, S. K. Tadikonda, and S. M. Collins, “Knowledge-basedinterpretation of mr brain images,” IEEE Transactions on MedicalImaging,, vol. 15, no. 4, pp. 443–452, 1996.
[25] S. Shen, W. Sandham, M. Granat, and A. Sterr, “Mri fuzzy seg-mentation of brain tissue using neighborhood attraction with neural-network optimization,” IEEE Transactions on Information Technologyin Biomedicine, vol. 9, no. 3, pp. 459–467, 2005.
[26] D. L. Pham, C. Xu, and J. L. Prince, “Current methods in medical imagesegmentation,” Annual Review of Biomedical Engineering, vol. 2, no. 1,pp. 315–337, 2000.
[27] P. Hore, L. O. Hall, D. B. Goldgof, Y. Gu, A. A. Maudsley, andA. Darkazanli, “A scalable framework for segmenting magnetic reso-nance images,” Journal of Signal Processing Systems, vol. 54, no. 1-3,pp. 183–203, 2008.
[28] J. Kang, L. Min, Q. Luan, X. Li, and J. Liu, “Novel modified fuzzy c-means algorithm with applications,” Digital Signal Processing, vol. 19,no. 2, pp. 309–319, 2009.
[29] A. K. Jain, M. N. Murty, and P. J. Flynn, “Data clustering: a review,”ACM Comput. Surv., vol. 31, no. 3, pp. 264–323, 1999.
[30] S. E. Brian, L. Sabine, and L. Morven. Wiley Publishing, 2009.[31] H. Zhou, “Fuzzy c-means and its applications in medical imaging,”
in Computational Intelligence in Medical Imaging: Techniques andApplications, 2009, p. 213.
[32] A. Rosenfeld and A. C. Kak, Digital Picture Processing. AcademicPress, Inc. New York, NY., 1982, vol. 2nd.
[33] A. K. Jain, R. P. W. Duin, and M. Jianchang, “Statistical patternrecognition: a review,” IEEE Transactions on Pattern Analysis andMachine Intelligence,, vol. 22, no. 1, pp. 4–37, 2000.
[34] O. M. Alia, R. Mandava, D. Ramachandram, and M. E. Aziz, “Harmonysearch- based cluster initialization for fuzzy c-means segmentation ofmr images.” in International Technical Conference of IEEE Region 10(TENCON), Singapore, 2009, pp. 1–6.
[35] O. M. Alia, R. Mandava, D. Ramachandram, and M. E. Aziz, “Anovel image segmentation algorithm based on harmony fuzzy searchalgorithm,” in International Conference of Soft Computing and PatternRecognition, SOCPAR ’09., Malacca, Malaysia, 2009, pp. 335–340.
[36] J. C. Bezdek and R. J. Hathaway, “Optimization of fuzzy clusteringcriteria using genetic algorithms,” in Proceedings of the First IEEEConference on Evolutionary Computation,IEEE World Congress onComputational Intelligence, vol. 2, Orlando, FL, USA, 1994, pp. 589–594.
[37] U. Maulik and I. Saha, “Modified differential evolution based fuzzyclustering for pixel classification in remote sensing imagery,” PatternRecognition, vol. 42, no. 9, pp. 2135–2149, 2009.
[38] P. M. Kanade and L. O. Hall, “Fuzzy ants and clustering,” IEEETransactions on Systems, Man and Cybernetics,Part A,, vol. 37, no. 5,pp. 758–769, 2007.
[39] K. S. Al-Sultan and C. A. Fedjki, “A tabu search-based algorithm forthe fuzzy clustering problem,” Pattern Recognition, vol. 30, no. 12, pp.2023–2030, 1997.
[40] S. Das, A. Abraham, and A. Konar, “Metaheuristic pattern clustering -an overview,” in Metaheuristic Clustering. Springer, 2009, pp. 1–62.
[41] S. Das and A. Konar, “Automatic image pixel clustering with animproved differential evolution,” Applied Soft Computing, vol. 9, no. 1,pp. 226–236, 2009.
[42] S. Saha and S. Bandyopadhyay, “A new point symmetry based fuzzygenetic clustering technique for automatic evolution of clusters,” Infor-mation Sciences, vol. 179, no. 19, pp. 3230–3246, 2009.
[43] S. Saha and S. Bandyopadhyay, “Mri brain image segmentation by fuzzysymmetry based genetic clustering technique,” in IEEE Congress onEvolutionary Computation, CEC, 2007, pp. 4417–4424.
[44] R. Campello, E. Hruschka, and V. Alves, “On the efficiency of evolution-ary fuzzy clustering,” Journal of Heuristics, vol. 15, no. 1, pp. 43–75,2009.
[45] M. K. Pakhira, S. Bandyopadhyay, and U. Maulik, “A study of somefuzzy cluster validity indices, genetic clustering and application to pixelclassification,” Fuzzy Sets and Systems, vol. 155, no. 2, pp. 191–214,2005.
[46] U. Maulik and S. Bandyopadhyay, “Fuzzy partitioning using a real-coded variable-length genetic algorithm for pixel classification,” IEEETransactions on Geoscience and Remote Sensing., vol. 41, no. 5, pp.1075–1081, 2003.
[47] E. R. Hruschka, R. J. G. B. Campello, A. A. Freitas, and A. C. P. L. F. d.Carvalho, “A survey of evolutionary algorithms for clustering,” IEEETransactions on Systems, Man and Cybernetics, Part C: Applicationsand Reviews, vol. 39, pp. 133–155, 2009.
[48] D. Horta, M. Naldi, R. Campello, E. Hruschka, and A. de Carvalho,“Evolutionary fuzzy clustering: An overview and efficiency issues,” inFoundations of Computational Intelligence, 2009, pp. 167–195.
[49] S. Das, A. Abraham, and A. Konar, Metaheuristic Clustering. SpringerVerlag, Germany, 2009.
[50] Z. W. Geem, J. H. Kim, and G. Loganathan, “A new heuristic opti-mization algorithm: harmony search.” SIMULATION, vol. 76, no. 2, pp.60–68, 2001.
[51] “Brain web: Simulated brain database. mcconnell brain imagingcentre. montreal neurological institute, mcgill university,http://www.bic.mni.mcgill.ca/brainweb.”
[52] J. C. Bezdek, Pattern Recognition with Fuzzy Objective Function Algo-rithms. Kluwer Academic Publishers, 1981.
[53] A. P. Dempster, N. M. Laird, and D. B. Rubin, “Maximum likelihoodfrom incomplete data via the em algorithm,” Journal of the RoyalStatistical Society. Series B (Methodological), vol. 39, no. 1, pp. 1–38,1977.
[54] G. Ingram and T. Zhang, in Music-Inspired Harmony Search Algorithm.Springer Berlin / Heidelberg, 2009, pp. 15–37.
[55] Z. W. Geem, Harmony Search Algorithms for Structural Design Opti-mization. Springer Berlin / Heidelberg, 2009.
[56] Z. W. Geem, C.-L. Tseng, and Y. Park, “Harmony search for generalizedorienteering problem: Best touring in china,” in Advances in NaturalComputation. Springer Berlin / Heidelberg, 2005, pp. 741–750.
[57] M. K. Pakhira, S. Bandyopadhyay, and U. Maulik, “Validity index forcrisp and fuzzy clusters,” Pattern Recognition, vol. 37, no. 3, pp. 487–501, 2004.
[58] A. Ben-Hur and I. Guyon, “Detecting stable clusters using principal com-ponent analysis,” Methods in Molecular Biology-Clifton then Totowa,vol. 224, pp. 159–182, 2003.