Fig. 2.

The structure of the network. The generator creates the synthetic ground glass nodule at the position where the mask in the input. The generator is composed of convolutional layers with a kernel size of 3 × 3, the batch normalisation, and the “parametric rectified linear unit” (PReLU) activation function. The discriminator was composed of convolutional layers with a kernel size of 3 × 3, the batch normalisation, and the leaky PReLU activation function