Table 3.

Sampling of the receptive field calculations at the output of every layer of the proposed network. The different types of layers are abbreviated with the following notation: I: input image, C: convolution block, M: max pooling, U: max unpooling, D: depth concatenation, O: output image.

Layer Type	I	C	C	M	C	⋯	C	M	U	C	⋯	C	U	D	C	O
l	1	2	3	4	5	⋯	18	19	20	21	⋯	37	38	39	40	41
$s_l$	1	1	1	2	1	⋯	1	2	0.5	1	⋯	1	0.5	1	1	1
$s_{l_{\mathrm{effective}}}$	1	1	1	2	2	⋯	16	32	16	16	⋯	2	1	1	1	1
$w_l$	1	3	3	1	3	⋯	3	1	1	3	⋯	3	1	1	3	3
$r_l$	1	3	5	5	9	⋯	181	181	181	213	⋯	359	359	359	361	363