MRI Brain Images Healthy and Pathological Tissues Classification with the Aid of Improved Particle Swarm Optimization and Neural Network

Table 2

Performance of the proposed improved particle swarm optimization and neural network (IPSONN)classification method in classifying (a) WM, (b) GM, (c) CSF, (d) edema, and (e) tumor from five different brain MRI images.

(a)  

Images	TP	FP	TN	FN	Sensitivity	FPR	ACC	Specificity	PPV	NPV	FDR	MCC
1	1	0	4	0	100	0.0	100	100	100	100	0	44.7
2	0	1	4	0	0	20.0	80	80	0	100	100	0
3	1	0	4	0	100	0.0	100	100	100	100	0	44.7

(b)  

Images	TP	FP	TN	FN	Sensitivity	FPR	ACC	Specificity	PPV	NPV	FDR	MCC
1	1	1	3	0	100	25.0	80	75	50	100	50	43.3
2	1	0	4	0	100	0.0	100	100	100	100	0	44.7
3	1	0	4	0	100	0.0	100	100	100	100	0	44.7

(c)  

Images	TP	FP	TN	FN	Sensitivity	FPR	ACC	Specificity	PPV	NPV	FDR	MCC
1	1	0	4	0	100	0.0	100	100	100	100	0	44.7
2	1	0	4	0	100	0.0	100	100	100	100	0	44.7
3	0	1	4	0	100	0.0	100	100	100	100	0	44.7

(d)  

Images	TP	FP	TN	FN	Sensitivity	FPR	ACC	Specificity	PPV	NPV	FDR	MCC
1	1	0	4	0	100	0.0	100	100	100	100	0	44.7
2	1	1	3	0	100	25.0	80	75	50	100	50	43.3
3	1	0	4	0	100	0.0	100	100	100	100	0	44.7

(e)  

Images	TP	FP	TN	FN	Sensitivity	FPR	ACC	Specificity	PPV	NPV	FDR	MCC
1	0	1	4	0	0	20.0	80	80	0	100	100	0
2	1	0	4	0	100	0.0	100	100	100	100	0	44.7
3	1	0	4	0	100	0.0	100	100	100	100	0	44.7