TABLE 1.

Comparison of mimicry-embedding performance applied to various neural network architectures^a

Condition	Precision	Recall	F1 score	AUC
Feed forward: data set	0.96	0.96	0.96	0.98
Feed forward: mimicry-embedded data set	0.96	0.96	0.96	0.98
LeNet: data set	0.79	0.79	0.76	0.87
LeNet: mimicry-embedded data set	0.81	0.81	0.79	0.89
ResNet-101: data set	0.92	0.92	0.92	0.96
ResNet-34: mimicry-embedded data set	0.96	0.96	0.96	0.98
CapsNet: data set
CapsNet: mimicry-embedded data set	0.96	0.96	0.96	0.98

^aAUC, area under the receiver operating characteristics curve. All metrics are averaged as one-versus-rest across classes. A missing value indicates “no convergence.”