Multisource DOA Estimation in Impulsive Noise Environments Using Convolutional Neural Networks

Table 5

DOA2 subnetwork architecture.

Layer and activation		Output shape	Parameter
L0:	Input	8×8	0
L1:	Convolution + Relu	64 × 8 × 8	640
L2:	Convolution + Relu	64 × 8 × 8	36928
L3:	Convolution + Relu	32 × 8 × 8	18464
L4:	Convolution + Relu	16 × 8 × 8	4624
L5:	Flatten	1024	0
L6:	Dense + Relu	10	10250
L7:	Dense + Relu	2	22