Mathematical Problems in Engineering

Research Article

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

Segmentation performance results for image “a” using GMM.


GMM	Cluster 1	Cluster 2	Cluster 3	Cluster 4	Metrics

					Best BAcc.: 0.87485 Mean BAcc: 0.76002 Mean sen.: 0.65702 Mean spe.: 0.82868 Avg. time: 7.4548

n					Best BAcc.: 0.93637 Mean BAcc: 0.82969 Mean sen.: 0.53766 Mean spe.: 0.88299 Avg. time: 5.14159

					Best BAcc.: 0.86040 Mean BAcc: 0.7717 Mean sen.: 0.58555 Mean spe.: 0.85133 Avg. time: 4.5137

					Best BAcc.: 0.95228 Mean BAcc: 0.84635 Mean sen.: 0.71681 Mean spe.: 0.90009 Avg. time: 3.1992

YCrCb					Best BAcc.: 0.79657 Mean BAcc: 0.75988 Mean sen.: 0.64811 Mean spe.: 0.83186 Avg. time: 4.7854

CrCb					Best BAcc.: 0.94294 Mean BAcc: 0.81556 Mean sen.: 0.61425 Mean spe.: 0.87505 Avg. time: 3.1574

					Best BAcc.: 0.91797 Mean BAcc: 0.85888 Mean sen.: 0.72094 Mean spe.: 0.90462 Avg. time: 4.2984

					Best BAcc.: 0.91816 Mean BAcc: 0.83095 Mean sen.: 0.50944 Mean spe.: 0.89749 Avg. time: 3.4293

Orig					Best BAcc.: 0.91808 Mean BAcc: 0.76817 Mean sen.: 0.22684 Mean spe.: 0.96200 Avg. time: 2.82860