Advances, Systems and Applications
Table 4 Summary of pre-trained CNN models used in this research
From: Security strategy for autonomous vehicle cyber-physical systems using transfer learning
Model Name | Depth | Number of Layers | Image Size | Replace Final Layer |
|---|---|---|---|---|
InceptionV3 | 48 | 316 | 299 X 299 | Fully connected layer, Classification layer |
ResNet-50 | 50 | 177 | 224 X 224 | Fully connected layer, Classification layer |
ShuffleNet | 50 | 173 | 224 X 224 | Fully connected layer, Classification layer |
MobileNetV2 | 53 | 155 | 224 X 224 | Fully connected layer, Classification layer |
GoogLeNet | 22 | 144 | 224 X 224 | Fully connected layer, Classification layer |
ResNet-18 | 18 | 72 | 224 X 224 | Fully connected layer, Classification layer |
SqueezeNet | 18 | 68 | 227 X 227 | Convolution2dLayer, Classification layer |
AlexNet | 8 | 25 | 227 X 227 | Fully connected layer, Classification layer |