Table 1 Parameter setup for training the model of neural network
From: Artifact suppression for breast specimen imaging in micro CBCT using deep learning
| Â | Modified U-Net in sinogram domain | Modified ResU-Net in image domain |
|---|---|---|
Number of image patches for training | 180,000 | 458,640 |
Number of image patches for validation | 18,000 | 80,000 |
Size of image patches | 64 × 64 | 48 × 48 |
Epoch | 100 | 1300 |
Optimizer | Adam | Adam |
Learning rate | 0.00001 | 0.001 |
Loss Function | RMSE | RMSE |