Table 7.

The classification testing accuracy (%) for the five classifiers used in MB-AI-His trained using spatial DL features extracted from the three CNNs.

Binary Classification Level
Features	Linear-SVM	Cubic-SVM	k-NN	Linear Discriminant Analysis(LDA)	Ensemble Subspace Discriminant(ESD)
Spatial-ResNet-50	99.6	100	98	100	94.8
Spatial-MobileNet	99.2	99.2	98.2	99	99.4
Spatial-DenseNet-201	98	98	97.6	98	99.2
Multi-Class Classification Level
Spatial-ResNet-50	93.66	94.96	88.18	95.74	93.9
Spatial-MobileNet	91.72	92.2	87.28	94.54	93.4
Spatial-DenseNet-201	94.32	96.26	92.88	97.16	94.84