Multimodal MRI Brain Tumor Image Segmentation Using Sparse Subspace Clustering Algorithm

Table 7

Comparison of multimodal image segmentation results with 15% noise.

Experimental sample		Index
		Dice	Jaccard	Precision	Recall
Malignant tumor	1	0.7010	0.6235	0.7788	0.6963
	2	0.7121	0.6417	0.8293	0.6625
	3	0.7006	0.6582	0.8126	0.6741
	4	0.7259	0.7128	0.8253	0.6701
	5	0.7361	0.6733	0.8256	0.6693
	6	0.7140	0.6682	0.7896	0.6642
	7	0.7263	0.7117	0.7526	0.6723
	8	0.7030	0.6336	0.7864	0.6008
	9	0.7234	0.6402	0.8124	0.6037
	10	0.7026	0.6513	0.8102	0.6395
	11	0.7263	0.6412	0.8006	0.6279
	12	0.6136	0.4562	0.8152	0.4963
	13	0.6742	0.5846	0.7852	0.6230
	14	0.6892	0.6006	0.7984	0.6172
	15	0.7211	0.6001	0.8010	0.6003
Bright tumor	1	0.6982	0.5895	0.7142	0.6110
	2	0.7120	0.7312	0.7265	0.6753
	3	0.7361	0.5742	0.7416	0.5996
	4	0.7216	0.5863	0.6246	0.7582
	5	0.7121	0.7323	0.7693	0.7296
	6	0.7132	0.7125	0.8263	0.7369
	7	0.6030	0.4852	0.8082	0.5020
	8	0.6482	0.5611	0.8060	0.5801
	9	0.7413	0.7230	0.8132	0.6778
	10	0.7143	0.6736	0.8007	0.6256
Mean		0.7028	0.6346	0.7862	0.6405