Automatic Diagnosis of Alzheimer’s Disease and Mild Cognitive Impairment Based on CNN + SVM Networks with End-to-End Training

Table 3

Evaluation of the proposed 3DCNN + SVM with E2E applied to binary classification of AD vs. NC samples (%).

Method	Training set				Testing set
	ACC	SEN	SPE	AUC	ACC	SEN	SPE	AUC
Gray [6]	—	—	—	—	81.60	82.7	80.4	90.0
Lu [10]	—	—	—	—	89.44	88.89	90.0	—
Silveira [8]	—	—	—	—	90.97	—	—	—
Ding et al. [12]	98.92	99.49	98.61	98.95	86.27	86.97	85.78	90.50
Liu et al. [14]	98.61	99.59	98.07	99.84	89.31	87.50	90.32	92.96
Huang et al. [15]	99.21	99.43	98.48	99.35	88.68	87.74	89.17	91.98
Proposed	99.19	99.39	99.54	99.88	90.82	91.29	90.59	93.75