Table 6 Results of FGSM-based adversarial testing on the OASIS dataset

From: Automated Alzheimer’s disease detection using active learning model with reinforcement learning and scope loss function

Model

Accuracy

F-measure

G-measure

AUC

HBOA-MLP19

68.699 ± 0.023

75.050 ± 0.032

76.073 ± 0.005

0.571 ± 0.060

CNN20

69.461 ± 0.049

76.616 ± 0.017

77.599 ± 0.016

0.587 ± 0.085

VGG1621

70.234 ± 0.034

77.658 ± 0.043

78.640 ± 0.056

0.604 ± 0.014

CLADSI22

71.238 ± 0.089

78.557 ± 0.063

79.613 ± 0.030

0.618 ± 0.090

DQN29

72.973 ± 0.027

79.865 ± 0.096

80.879 ± 0.079

0.630 ± 0.051

4D-AlzNet50

74.320 ± 0.084

80.883 ± 0.098

81.948 ± 0.053

0.639 ± 0.063

XAI51

74.977 ± 0.038

81.886 ± 0.061

82.917 ± 0.004

0.657 ± 0.055

RESNET50 KNN52

75.549 ± 0.090

83.252 ± 0.028

84.224 ± 0.039

0.672 ± 0.056

Proposed

83.838 ± 0.022

90.087 ± 0.078

90.957 ± 0.071

0.735 ± 0.039

Back to article page