Research Article

Improved Unsupervised Color Segmentation Using a Modified Color Model and a Bagging Procedure in -Means++ Algorithm

Table 12

Segmentation performance results for image “f” using GMM.

GMMCluster 1Cluster 2Cluster 3Cluster 4Metrics

Best BAcc.: 0.99634
Mean BAcc: 0.93741
Mean sen.: 0.80034
Mean spe.: 0.95122
Avg. time: 5.87

nBest BAcc.: 0.98807
Mean BAcc: 0.86814
Mean sen.: 0.49632
Mean spe.: 0.84997
Avg. time: 3.437

Best BAcc.: 0.99766
Mean BAcc: 0.95731
Mean sen.: 0.86890
Mean spe.: 0.95453
Avg. time: 3.38859

Best BAcc.: 0.99832
Mean BAcc: 0.91859
Mean sen.: 0.69350
Mean spe.: 0.90564
Avg. time: 2.8533

YCrCbBest BAcc.: 0.99738
Mean BAcc: 0.94940
Mean sen.: 0.84076
Mean spe.: 0.95318
Avg. time: 3.4557

CrCbBest BAcc.: 0.99872
Mean BAcc: 0.92172
Mean sen.: 0.76444
Mean spe.: 0.91761
Avg. time: 2.2925

Best BAcc.: 0.93728
Mean BAcc: 0.88837
Mean sen.: 0.68992
Mean spe.: 0.93274
Avg. time: 4.0777

Best BAcc.: 0.97601
Mean BAcc: 0.89289
Mean sen.: 0.59888
Mean spe.: 0.94690
Avg. time: 2.65219

OrigBest BAcc.: 0.98304
Mean BAcc: 0.91677
Mean sen.: 0.75014
Mean spe.: 0.927527
Avg. time: 1.69729