Table 3 CNN configurations.

From: ASD-SWNet: a novel shared-weight feature extraction and classification network for autism spectrum disorder diagnosis

Layer

Settings

Output shape

Input

 

(2000)

Conv 0

Kernel size: 7

Filters: 16

(1994, 16)

MaxPooling 0

Pooling window: 2

(997, 16)

Conv 1

Kernel size: 5

Filters: 32

(993, 32)

MaxPooling 1

Pooling window: 2

(496, 32)

Conv 2

Kernel size: 3

Filters: 64

(494, 64)

Conv 3

Kernel size: 3

Filters: 128

(492, 128)

MaxPooling 2

Pooling window: 2

(246, 128)

Conv 4

Kernel size: 3

Filters: 256

(244, 256)

Conv 5

Kernel size: 3

Filters: 512

(242, 512)

Conv 6

Kernel size: 3

Filters: 1024

(240, 1024)

MaxPooling 3

Pooling window: 2

(120, 1024)

Fully connected

 

(256)

Sigmoid

 

(1)

Back to article page