BCE evolution over the first 30 epochs in training (continued lines) and validation (dashed lines) steps. UNet++[10⁻²] schemes is removed as its corresponding BCE diverged. The displayed schemes show similar performances. Models with a learning-rate value of 10⁻⁵ resulted in higher BCEs. The lowest error is reported for UNet++[10⁻³] (purple), and the highest error corresponds to the UNet[10⁻⁵] scheme (red). A plateau regime is reached within the first ten epochs for all models, which confirms the rapid convergence of the UNet-based networks.