Table 3 Decoder architecture of the adaptive physics autoencoder, using three transpose Convolution layers, followed by a final Convolutional layer with a sigmoid activation function.
From: Enhanced confocal microscopy with physics-guided autoencoders via synthetic noise modeling
Layer | Type | Input channels | Output channels | Kernel size | Stride | Padding |
|---|---|---|---|---|---|---|
1 | Transpose Conv + LeakyReLU | 512 | 256 | 3 × 3 | 2 | 1 |
2 | Transpose Conv + LeakyReLU | 256 | 128 | 3 × 3 | 2 | 1 |
3 | Transpose Conv + LeakyReLU | 128 | 64 | 3 × 3 | 2 | 1 |
4 | Conv + sigmoid | 64 | 1 | 3 × 3 | 1 | - |
