• Laser & Optoelectronics Progress
  • Vol. 59, Issue 22, 2210001 (2022)
Chao Wang1,*, Yongshun Wang1, and Fan Di2
Author Affiliations
  • 1School of Electronic and Information Engineering, Lanzhou Jiaotong University, Lanzhou 730070, Gansu , China
  • 2Diaoyutai Hotel Administration, Ministry of Foreign Affairs, Beijing 100080, China
  • show less
    DOI: 10.3788/LOP202259.2210001 Cite this Article Set citation alerts
    Chao Wang, Yongshun Wang, Fan Di. Fast and Automatic Fuzzy C-Means Clustering Color Image Segmentation Algorithm[J]. Laser & Optoelectronics Progress, 2022, 59(22): 2210001 Copy Citation Text show less
    Overall block diagram of proposed algorithm
    Fig. 1. Overall block diagram of proposed algorithm
    Comparison of different superpixel algorithms. (a) Original image; (b) LSC; (c) Mean shift; (d) WT; (e) MMGR-WT; (f) proposed algorithm
    Fig. 2. Comparison of different superpixel algorithms. (a) Original image; (b) LSC; (c) Mean shift; (d) WT; (e) MMGR-WT; (f) proposed algorithm
    Color space, histogram, and number of clusters of images. (a) Original images; (b) original image color space;(c) superpixel image color space; (d) original image histogram; (e) superpixel image histogram; (f) number of clusters
    Fig. 3. Color space, histogram, and number of clusters of images. (a) Original images; (b) original image color space;(c) superpixel image color space; (d) original image histogram; (e) superpixel image histogram; (f) number of clusters
    Segmentation results of 5 algorithms on BSDS500 database. (a) Original images; (b) FCM; (c) SFFCM; (d) FCM_SICM;(e) AFCF; (f) proposed method
    Fig. 4. Segmentation results of 5 algorithms on BSDS500 database. (a) Original images; (b) FCM; (c) SFFCM; (d) FCM_SICM;(e) AFCF; (f) proposed method
    Segmentation results of proposed algorithm on AID and MSRC databases. (a) Original images; (b) superpixel segmentation;(c) segmentation results of proposed algorithm
    Fig. 5. Segmentation results of proposed algorithm on AID and MSRC databases. (a) Original images; (b) superpixel segmentation;(c) segmentation results of proposed algorithm
    Segmentation indexSA /%VPC /%VPE /%Running time /s
    FCM92.7892.6523.2717
    FCM_SICM90.2591.8925.5222
    SFFCM89.8890.1319.8916
    AFCF89.9091.9217.7228
    Proposed method93.9292.8013.4320
    Table 1. Five algorithms for average segmentation index of color images
    AlgorithmCalculation step ECalculation step function E(n)Time complexity
    FCMH×W×K×Tn4O(n4)
    FCM_SICMH×W×K×Tn4O(n4)
    SFFCMH×W×h2×TM+TW×N×K×Tn5+n4O(n5)
    AFCFN×K×Tn3O(n3)
    Proposed methodN×K×Tn3O(n3)
    Table 2. 5 algorithm time complexity