Table 6.

Hyperparameters considered for the learning algorithms.

Hyperparameters	CNN-MNIST	CNN-Angio	FCNN-WBC
Objective function	CE	BCE	MSE
Weight initialization method	Xavier	Xavier	Xavier
Optimizer		Minibatch gradient descent
Momentum	0.9	0.9	0.9
Learning rate	0.01	0.01	0.01
Batch size	32	16	32
Dropout rate	—	—	25%
Epochs	100	100	4500