Multimodal MRI Brain Tumor Image Segmentation Using Sparse Subspace Clustering Algorithm

Table 6

Comparison of multimodal image segmentation results with 10% noise.

Experimental sample		Index
		Dice	Jaccard	Precision	Recall
Malignant tumor	1	0.7585	0.7147	0.8452	0.7581
	2	0.7662	0.7020	0.8967	0.7052
	3	0.7596	0.6996	0.8746	0.7196
	4	0.8008	0.7989	0.9003	0.8320
	5	0.8020	0.7011	0.9114	0.7404
	6	0.7642	0.7102	0.9095	0.7462
	7	0.8001	0.7834	0.8482	0.8346
	8	0.7779	0.6996	0.9063	0.6730
	9	0.7823	0.7032	0.9101	0.7162
	10	0.7863	0.7142	0.9011	0.7285
	11	0.7903	0.7020	0.9039	0.7126
	12	0.6523	0.5011	0.9008	0.5028
	13	0.7124	0.6031	0.8557	0.6896
	14	0.7210	0.6313	0.8932	0.6745
	15	0.7695	0.6220	0.8712	0.6512
Bright tumor	1	0.7103	0.6102	0.8103	0.6326
	2	0.7533	0.7945	0.8410	0.8071
	3	0.8120	0.6103	0.8124	0.6426
	4	0.8236	0.6120	0.6731	0.8426
	5	0.8022	0.8008	0.8526	0.7945
	6	0.8471	0.8106	0.8989	0.8142
	7	0.6326	0.5030	0.9009	0.5231
	8	0.7002	0.5936	0.8855	0.6005
	9	0.8308	0.7852	0.8797	0.8030
	10	0.8001	0.7262	0.8722	0.7103
Mean		0.7662	0.6853	0.8702	0.7142