Skip to main content

Table 2 Hyperparameters used for training the DNNs

From: COVID-19 lung CT image segmentation using deep learning methods: U-Net versus SegNet

Exp. SegNet U-NET
ILR MniBatch ILR MiniBatch
1 \({1}{\mathrm{e}}{-4}\) 4 \({1}{\mathrm{e}}{-4}\) 2
2 \({1}{\mathrm{e}}{-4}\) 8 \({1}{\mathrm{e}}{-4}\) 8
3 \({1}{\mathrm{e}}{-4}\) 12 \({1}{\mathrm{e}}{-4}\) 12
4 \({1}{\mathrm{e}}{-3}\) 4 \({5}{\mathrm{e}}{-4}\) 2
5 \({1}{\mathrm{e}}{-3}\) 8 \({5}{\mathrm{e}}{-4}\) 8
6 \({1}{\mathrm{e}}{-3}\) 12 \({5}{\mathrm{e}}{-4}\) 12
7 \({3}{\mathrm{e}}{-3}\) 4 \({1}{\mathrm{e}}{-3}\) 2
8 \({3}{\mathrm{e}}{-3}\) 8 \({1}{\mathrm{e}}{-3}\) 8
9 \({3}{\mathrm{e}}{-3}\) 12 \({1}{\mathrm{e}}{-3}\) 12
  1. Nine experiments for each network with different initial learning rates (ILR) and mini batch sizes