Table 2.

Summary of hyperparameters and their values used during the training process.

Hyper-parameter	Value
Type of input images	Grayscale
Resolution of input images	200 $\times$ 200
Learning rate optimizer	SGD optimizer with Cosine annealing scheduler
Learning rate	${10}^{-2}$ to ${10}^{-5}$
Training epochs	500 epochs
Snapshot interval	100 epochs
Batch size during training	16
Residual unit depth (d)	2