Fig 1

Confusion matrix diagram for introducing classification confidence"

Fig 2

Classification confidence centroid skewing quadrangle before and after adversarial attack"

Fig 3

Visual evaluation process of adversarial robustness"

Table 1

Basic information of the testing dataset"

Table 2

Visual evaluation index based on classification confidence-based confusion matrix"

Fig 4

Correlation diagram of $ {\boldsymbol{\Delta}} {{\boldsymbol{S}}} _{\bf{conf}} $ and $ {\boldsymbol{\Delta}} {{\bf{acc}}}_{\bf{conf}} $"

Table 3

Different $ {\bf{esp}} $ corresponding to ${\boldsymbol{ \Delta}} {{\boldsymbol{S}}}_{\bf{conf}} $ and ${\boldsymbol{ \Delta}} {{\bf{acc}}}_{\bf{conf}} $"

Fig 5

Sensitivity analysis diagram of $ {\boldsymbol{\Delta}} {{\boldsymbol{S}}}_\bf{conf} $ and $ {\boldsymbol{\Delta}} {{\bf{acc}}}_\bf{conf} $"

Table 4

$ {\boldsymbol{\Delta}} {{\boldsymbol{S}}} _{\bf{conf}} $ and $ {\boldsymbol{\Delta}} {{\bf{acc}}}_{\bf{conf}} $ of the cancer dataset"

Fig 6

Visual indicator of cancer dataset based on ANN classification algorithm (0.483 vs. 0.993 vs. 0.413)"

Fig 7

Visual indicator of cancer dataset based on LR classification algorithm (0.433 vs. 0.437 vs. 0.462)"

Fig 8

Visual indicator of cancer dataset based on SVM classification algorithm (0.444 vs. 0.451 vs. 0.417)"

Fig 9

Visual indicator of cancer dataset based on CNN classification algorithm (0.495 vs. 0.588 vs. 0.487)"

Fig 10

Visual indicator of cancer dataset based on Transformer classification algorithm (0.488 vs. 0.682 vs. 0.458)"