Local and Deep Features Based Convolutional Neural Network Frameworks for Brain MRI Anomaly Detection

Table 2

Parameters of DAE for brain anomalies.

Parameters	Values
Size of nods	[1200, 800, 400,200]
Number of layers	4
Pretraining learning rate	0.1
Pretraining epochs	10000
Minibatch size pretraining	32
Finetuning learning rate	0.001
Finetuning epochs	1000
Minibatch size finetuning	32
Loss function	Binary cross-entropy
Optimization	Adam
Number of epochs for training	1000
Number of epochs for finetunning	500