An Efficient Multi-Scale Local Binary Fitting-Based Level Set Method for Inhomogeneous Image Segmentation

Table 1

Region overlap metrics of different algorithms, where “image number 1” to “image number 6” are the serial numbers of the images shown in the first column of Figure 3 with the same order.

Input image	Algorithm	Indexes
		JS	DSC	FPR	FNR
Image number 1	CV	0.4705	0.6399	0.0450	0.5189
	GAC	0.7111	0.8311	0.1192	0.2132
	EM	0.5101	0.6756	0.1170	0.4529
	OTSU	0.7141	0.8332	0.1952	0.1363
	PCNN	0.7132	0.8326	0.2099	0.1200
	MLBF	0.9302	0.9638	0.0381	0.0343
Image number 2	CV	0.4818	0.6503	0.0147	0.5147
	GAC	0.3590	0.5283	0.0989	0.6263
	EM	0.8096	0.8948	0.1417	0.0655
	OTSU	0.0951	0.1736	0.0118	0.9048
	PCNN	0.1063	0.1922	0.0217	0.8935
	MLBF	0.8319	0.9082	0.0138	0.0539
Image number 3	CV	0.9035	0.9493	0.0453	0.0560
	GAC	0.6388	0.7796	0.1233	0.2982
	EM	0.9024	0.9487	0.0291	0.0724
	OTSU	0.8426	0.9146	0.1300	0.0360
	PCNN	0.8800	0.9362	0.0850	0.0416
	MLBF	0.9301	0.9638	0.0127	0.0286
Image number 4	CV	0.7423	0.8521	0.0403	0.2338
	GAC	0.5879	0.7405	0.0939	0.3740
	EM	0.7288	0.8431	0.0697	0.2291
	OTSU	0.7782	0.8753	0.1119	0.1371
	PCNN	0.7788	0.8757	0.1370	0.1113
	MLBF	0.8097	0.8948	0.0332	0.0385
Image number 5	CV	0.4534	0.6239	0.0444	0.5368
	GAC	0.6863	0.8140	0.1038	0.2545
	EM	0.3488	0.5172	0.6509	0.0021
	OTSU	0.0750	0.1395	0.0107	0.9249
	PCNN	0.0740	0.1378	0.0158	0.9259
	MLBF	0.8330	0.9089	0.0103	0.0020
Image number 6	CV	0.8697	0.9303	0.0199	0.1147
	GAC	0.7348	0.8471	0.0002	0.2651
	EM	0.8707	0.9309	0.0249	0.1095
	OTSU	0.8923	0.9431	0.0375	0.0756
	PCNN	0.8729	0.9321	0.0411	0.0931
	MLBF	0.9005	0.9477	0.0001	0.0133