Skip to main content
. 2019 Feb 25;19(4):972. doi: 10.3390/s19040972

Table 2.

Details of AICNN-1D.

Layer Kernel Size Kernel Number Stride Activation Function Output Size Padding
Conv_1 15 16 7 ReLU 286 × 16 Yes
Max-pooling 1 2 16 2 None 143 × 16 Yes
Conv_2 10 32 5 ReLU 29 × 32 Yes
Max-pooling 2 2 32 2 None 15 × 32 Yes
Conv_3 5 H1 2 ReLU 8 × H Yes
Global average pooling 8 H 1 None H No

1 H represents the number of classification labels.