From: Signal processing of Internet of Vehicles based on intelligent interference
Network structure | Convolution kernel structure | Output layer |
---|---|---|
Input | / | (2, 1024) |
Convolutional layer (normalization) | (2, 2, 128) | / |
ReLU/PReLU | 128 | / |
Convolutional layer (normalization) | (1, 2, 128) | / |
ReLU/PReLU | 128 | / |
Pooling layer (max pooling) | (1, 2, 128) stride = 2 | / |
Convolutional layer (normalization) | (1, 2, 256) | / |
ReLU/PReLU | 256 | / |
Convolutional layer (normalization) | (1, 2, 256) | / |
ReLU/PReLU | 256 | / |
Pooling layer (max pooling) | (1, 2, 256) stride = 2 | / |
Convolutional layer (normalization) | (1, 2, 512) | / |
ReLU/PReLU | 512 | / |
Convolutional layer (normalization) | (1, 2, 512) | / |
ReLU/PReLU | 512 | / |
Pooling layer (max pooling) | (1, 2, 512) stride = 2 | / |
Fully connected layer (dropout = 0.5) | 256*128*1 | 1000*1 |
Fully connected layer (dropout = 0.5) | 1000*1 | 1000*1 |
Classifier | 1000*3 | 3*1 |