Table 10 TPR comparison of different models with and without LHE_Gabor, using malGAN and DCGAN configurations.

From: Mobile malware detection method using improved GhostNetV2 with image enhancement technique

Detection model

RGB image

RGB image with LHE_Gabor

 

MalGAN

DCGAN

MalGAN

DCGAN

ReNet34

0.8196

0.8000

0.9031

0.8981

ResNet50

0.8717

0.8651

0.9142

0.9094

ResNet101

0.8609

0.8554

0.8951

0.8864

ResNet152

0.8041

0.7984

0.9164

0.9089

MobileNetV2

0.7858

0.7751

0.8382

0.8251

MobileNetV3

0.8096

0.7982

0.9014

0.8969

DenseNet121

0.8517

0.8451

0.9021

0.8868

DenseNet169

0.8668

0.8893

0.9172

0.9039

DenseNet201

0.8638

0.8692

0.9030

0.8981

ShuffleNetV2

0.7036

0.6951

0.8460

0.8391

ESPNetV2

0.7314

0.7180

0.8412

0.8361

EfficieNetV2

0.8534

0.8680

0.9082

0.9042

EfficienNetb0

0.7731

0.7600

0.9124

0.9072

EfficienNetb1

0.8563

0.8482

0.9162

0.9114

EfficienNetb2

0.8484

0.8414

0.8851

0.8600

EfficienNetb3

0.8591

0.8535

0.9142

0.9082

EfficienNetb4

0.8621

0.8572

0.9030

0.9014

EfficienNetb5

0.8591

0.8500

0.8931

0.8890

GhostNetV2

0.8314

0.8241

0.8921

0.8875

Proposed

0.8537

0.8482

0.9200

0.9164

Back to article page