Research Article
A Novel Malware Detection and Family Classification Scheme for IoT Based on DEAM and DenseNet
Table 2
The structure of DenseNet in the proposed model and its comparison with DenseNet-121.
| | Layers | Our DenseNet (k = 12) | DenseNet-121 (k = 32) |
| | Convolution | conv, stride 1 | conv, stride 2 | | Pooling | max pool, stride 2 | max pool, stride 2 | | DenseBlock(1) | | | | Transition | 11 conv | | | ave pool, stride 2 | | DenseBlock(2) | | | | Transition | conv | | | ave pool, stride 2 | | DenseBlock(3) | | | | Transition | 11 conv | | | ave pool, stride 2 | | DenseBlock(4) | | | | Pooling | GlobalAveragePooling | | | Softmax |
|
|
