Skip to main content
. 2022 Feb 16;22(4):1521. doi: 10.3390/s22041521

Table 5.

FDC-FS Model Hyper-Parameters.

Layer (Type) Output Shape Param #
Input (None, 128, 128, 1) 0
Conv2D (None, 128, 128, 32) 320
BatchNormalization (None, 128, 128, 32) 128
Conv2D (None, 126, 126, 32) 9248
BatchNormalization (None, 126, 126, 32) 128
MaxPooling2D (None, 63, 63, 32) 0
DropOut (None, 63, 63, 32) 0
Conv2D (None, 63, 63, 64) 18,496
BatchNormalization (None, 63, 63, 64) 256
Conv2D (None, 61, 61, 64) 36,928
BatchNormalization (None, 61, 61, 64) 256
MaxPooling2D (None, 31, 31, 64) 0
DropOut (None, 31, 31, 64) 0
Conv2D (None, 124, 124, 24) 624
Conv2D (None, 31, 31, 64) 36,928
MaxPooling2D (None, 31, 62, 24) 0
BatchNormalization (None, 31, 31, 64) 256
Activation (None, 31, 62, 24) 0
Conv2D (None, 126, 126, 32) 320
Conv2D (None, 29, 29, 64) 36,928
Conv2D (None, 27, 58, 48) 28,848
Activation (None, 126, 126, 32) 0
BatchNormalization (None, 29, 29, 64) 256
MaxPooling2D (None, 6, 29, 48) 0
MaxPooling2D (None, 31, 63, 32) 0
MaxPooling2D (None, 15, 15, 64) 0
Activation (None, 6, 29, 48) 0
Conv2D (None, 29, 61, 64) 18,496
DropOut (None, 15, 15, 64) 0
Conv2D (None, 2, 25, 48) 57,648
Activation (None, 29, 61, 64) 0
Conv2D (None, 15, 15, 128) 73,856
Activation (None, 2, 25, 48) 0
MaxPooling2D (None, 7, 30, 64) 0
BatchNormalization (None, 15, 15, 128) 512
Flatten (None, 2400) 0
Flatten (None, 13,440) 0
Conv2D (None, 13, 13, 128) 147,584
DropOut (None, 2400) 0
DropOut (None, 13,440) 0
BatchNormalization (None, 13, 13, 128) 512
Dense (None, 64) 153,664
Dense (None, 64) 860,224
MaxPooling2D (None, 7,7, 128) 0
Activation (None, 64) 0
Activation (None, 64) 0
DropOut (None, 7, 7, 128) 0
DropOut (None, 64) 0
DropOut (None, 64) 0
GlobalAveragePoolinh2D (None, 128) 0
Concatenate (None, 256) 0
Dense (None, 6) 1542
Total Parameters: 1,483,958
Trainable Parameters: 1,482,806
Non-Trainable Parameters: 1152