Table 1.

Details of the proposed RNN-WDCNN model structure.

Convolutional Pathway
No.	Layer Type	Kernel Size	Stride	Number of Kernels
C1	1D Convolution	256 × 1	2 × 1	16
C2	Max Pooling 1D	2 × 1	2 × 1	16
C3	1D Convolution	3 × 1	2 × 1	32
C4	Max Pooling 1D	2 × 1	2 × 1	32
C5	1D Convolution	3 × 1	2 × 1	64
C6	Max Pooling 1D	2 × 1	2 × 1	64
C7	1D Convolution	3 × 1	2 × 1	64
C8	Max Pooling 1D	2 × 1	2 × 1	64
C9	1D Convolution	3 × 1	2 × 1	64
C10	Max Pooling 1D	2 × 1	2 × 1	64
C11	Fully connected	100	N/A	1
Recurrent Pathway
No.	Layer Type	Kernel Size	Stride	Number of Kernels
R1	1D Convolution	256 × 1	2 × 1	16
R2	Max Pooling 1D	2 × 1	2 × 1	16
R3	RNN block	N/A	N/A	128
Output
No.	Layer Type	Kernel Size	Stride	Number of Kernels
O1	Softmax	10	N/A	1