An adaptive clustering segmentation algorithm based on FCM

The cluster number and the initial clustering centers must be reasonably set before the analysis of clustering in most cases. Traditional clustering segmentation algorithms have many shortcomings, such as high reliance on the specially established initial clustering center, tendency to fall into the local maximum point, and poor performance with multithreshold values. To overcome these defects, an adaptive fuzzy C-means segmentation algorithm based on a histogram (AFCMH), which synthesizes both main peaks of the histogram and optimized Otsu criterion, is proposed. First, the main peaks of the histogram are chosen by operations like histogram smoothing, merging of adjacent peaks, and ltering of small peaks, and then the values of main peaks are calculated. Second, a new separability measure is de ned and a group of main peaks with the maximum value of serve as the optimal segmentation threshold value. The values of these main peaks are employed for initializing of the initial clustering center. Finally, the image is segmented by the weighted fuzzy C-means clustering algorithm. The experiment results show that, compared with existing algorithms, the proposed method not only avoids the oversegmentation phenomenon but also has a signi cantly shorter computing time than the traditional segmentation algorithm based on mean shift. Therefore, the proposed algorithm can obtain satisfactory results and effectively improve executive efficiency.

PDF

___

[1] Mesejo P, Ibanez O, Cordon O, Cagnoni S. A survey on image segmentation using metaheuristic-based deformable models: state of the art and critical analysis. Appl Soft Comput 2016; 44: 1-29.
[2] Chen SC, Zhang DQ. Robust image segmentation using FCM with spatial constraints based on new kernel-induced distance measure. IEEE T Syst Man Cyb 2004; 34: 1907-1916.
[3] Nalwa VS, Binford TO. On detecting edges. IEEE T Pattern Anal 1986; 8: 699-714.
[4] Long JW, Shen XJ, Chen HP. Adaptive minimum error thresholding algorithm. Journal of Acta Automatica Sinica 2012; 38: 1134-1144 (in Chinese with abstract in English).
[5] Wang XH, Wan Y, Li R, Wang JL, Fang LL. A multi-object image segmentation C-V model based on region division and gradient guide. J Vis Commun Image R 2016; 39: 100-106.
[6] Zheng C, Qin Q, Liu G, Hu Y. Image segmentation based on multiresolution Markov random eld with fuzzy constraint in wavelet domain. IET Image Process 2012; 6: 213-221.
[7] Pal NR, Pal SK. A review on image segmentation techniques. Pattern Recogn 1993; 26: 1277-1294.
[8] Bezdek JC. Pattern Recognition with Fuzzy Objective Function Algorithms. New York, NY, USA: Plenum Press, 1981.
[9] Boujemaa N, Rocquencourt I. Generalized competitive clustering for image segmentation. In: 19th International Conference of the North American on Fuzzy Information Processing Society; 13{15 July 2000; Atlanta, GA, USA. New York, NY, USA: IEEE. pp. 133-137.
[10] Comaniciu D, Meer P. Mean shift: a robust approach toward feature analysis. IEEE T Pattern Anal 2002; 24: 603-619.
[11] Wang S, Xia Y, Jiao LC. Mean shift based adaptive texture image segmentation method. Journal of Software 2010; 21: 1451-1461 (in Chinese with abstract in English).
[12] Ye XQ, Huang ZH, Xiao Q. Histogram based fuzzy C-mean algorithm for image segmentation. In: Proceedings of the International Conference on Image, Speech and Signal Analysis; 7{9 May 1992; Tianjin, P. R. China. New York, NY, USA: IEEE. pp. 704-707.
[13] Liu JZ. A fuzzy clustering method for image segmentation based on two-dimensional histogram. Acta Electronica Sinica 1992; 9: 735-738 (in Chinese with abstract in English).
[14] Zhou X, Shen Q, Liu L. New two-dimensional fuzzy C-means clustering algorithm for image segmentation. Journal of Central South University of Technology 2008; 15: 882-887 (in Chinese with abstract in English).
[15] Lhoussaine M, Rachid Z, Ansari MEL. Image Segmentation. Rijeka, Croatia: InTech Open Access Publisher, 2011.
[16] Wang XZ, Wang YD, Wang LJ. Improving fuzzy C-means clustering based on feature-weight learning. Pattern Recogn 2004; 25: 1123-1132.
[17] Li D, Gu H, Zhang LY. A fuzzy C-means algorithm with interval-supervised attribute weights. Journal of Control and Decision 2010; 25: 456-460 (in Chinese with abstract in English).
[18] Jin HL, Zhu WP, Li LY. The best thresholding on 2-D gray level histogram. Journal of Pattern Recognition and Arti cial Intelligence 1999; 12: 329-333 (in Chinese with abstract in English).
[19] Wang WN, Wang HY, Mu WY. A novel histogram threshold auto-detection method. Journal of Computer Engi- neering and Applications 2005; 41: 89-90 (in Chinese with abstract in English).
[20] Otsu N. A threshold selection method from gray-level histograms. IEEE T Syst Man Cyb 1979; 9: 62-66.
[21] Zhao ZX, Cheng LZ, Cheng G. Neighborhood weighted fuzzy C-means clustering algorithm for image segmentation. IET Image Process 2014; 8: 150-161.
[22] Ahmed MN, Yamany SM, Mohamed N, Farag AA, Moriarty T. A modi ed fuzzy C-means algorithm for bias eld estimation and segmentation of MRI data. IEEE T Med Imaging 2002; 21: 193-199.
[23] Bezdek JC. Cluster validity with fuzzy sets. J Cybernetics 1974; 3: 58-73.
[24] Bezdek JC. Mathematical models for systematics and taxonomy. In: Proceedings of Eighth International Conference on Numerical Taxonomy; 10{12 August 1974; Oeiras, Portugal. San Francisco, CA, USA: W. H. Freeman. pp. 143- 166.