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

Table 6

Segmentation performance results for image “c” using GMM.

GMM	Cluster 1	Cluster 2	Cluster 3	Cluster 4	Cluster 5	Metrics
						Best BAcc.: 0.97231; Mean BAcc: 0.86137; Mean sen.: 0.64875; Mean spe.: 0.91404; Avg. time: 7.4137
n						Best BAcc.: 0.95474; Mean BAcc: 0.89090; Mean sen.: 0.50531; Mean spe.: 0.94607; Avg. time: 5.039
						Best BAcc.: 0.97013; Mean BAcc: 0.86512; Mean sen.: 0.65040; Mean spe.: 0.91608; Avg. time: 4.7831
						Best BAcc.: 0.98512; Mean BAcc: 0.96448; Mean sen.: 0.90266; Mean spe.: 0.97358; Avg. time: 3.16470
YCrCb						Best BAcc.: 0.97427; Mean BAcc: 0.86622; Mean sen.: 0.66280; Mean spe.: 0.91694; Avg. time: 5.13540
CrCb						Best BAcc.: 0.98513; Mean BAcc: 0.94658; Mean sen.: 0.88602; Mean spe.: 0.95928; Avg. time: 1.9721
						Best BAcc.: 0.98307; Mean BAcc: 0.88245; Mean sen.: 0.52690; Mean spe.: 0.94644; Avg. time: 5.60439
						Best BAcc.: 0.98099; Mean BAcc: 0.87426; Mean sen.: 0.55223; Mean spe.: 0.92978; Avg. time: 4.122
Orig						Best BAcc.: 0.96963; Mean BAcc: 0.92605; Mean sen.: 0.73088; Mean spe.: 0.94629; Avg. time: 1.90460