Multimodal MRI Brain Tumor Image Segmentation Using Sparse Subspace Clustering Algorithm

Table 8

Comparison of multimodal image segmentation results with 20% noise.

Experimental sample		Index
		Dice	Jaccard	Precision	Recall
Malignant tumor	1	0.6013	0.5582	0.5786	0.6030
	2	0.6230	0.5631	0.6023	0.5436
	3	0.6058	0.5477	0.6113	0.5633
	4	0.6003	0.5693	0.6037	0.5721
	5	0.6146	0.5746	0.6012	0.5126
	6	0.6008	0.5832	0.5963	0.5362
	7	0.6001	0.5746	0.5748	0.5284
	8	0.6200	0.5365	0.5836	0.5369
	9	0.5963	0.5208	0.5862	0.5623
	10	0.5982	0.5300	0.5916	0.5123
	11	0.5996	0.5613	0.5746	0.5023
	12	0.5342	0.4203	0.5842	0.4523
	13	0.5846	0.5110	0.5532	0.5203
	14	0.5768	0.5030	0.5631	0.5417
	15	0.5636	0.5007	0.5711	0.5731
Bright tumor	1	0.5875	0.5114	0.5369	0.5324
	2	0.5939	0.5630	0.5284	0.6064
	3	0.5742	0.5023	0.5748	0.5412
	4	0.5936	0.5431	0.5303	0.6234
	5	0.5768	0.5623	0.5923	0.6127
	6	0.5693	0.5665	0.5830	0.6471
	7	0.5234	0.4528	0.5746	0.4864
	8	0.5236	0.5220	0.5822	0.5520
	9	0.5741	0.5360	0.5623	0.6113
	10	0.5698	0.5142	0.5722	0.5436
Mean		0.5842	0.5331	0.5765	0.5527