Deep tomography for the three-dimensional atmospheric turbulence wavefront aberration

Home

All issues

Volume 687 (July 2024)

A&A, 687 (2024) A182

Full HTML

Open Access

Table 2

Detailed parameter configuration of the SF-ResNet architecture.

Layer	Type	Kernel Size	Stride	Padding	Input Size	Output Size
1	Conv + BN +FReLU	3 × 3 × 64	2	1	6× 360× 360	64 × 360 × 360
2	Max pooling	3 × 3	2	0	64 × 360 × 360	64 × 179 × 179
3	ResBlock + BN +FReLU	$[\begin{matrix} 1 \times 1 \times 64 \\ 3 \times 3 \times 64 \\ 1 \times 1 \times 256 \end{matrix}]$ $\[\left[\begin{array}{c}1 \times 1 \times 64 \\3 \times 3 \times 64 \\1 \times 1 \times 256\end{array}\right]\]$	$[\begin{array}{l} 1 \\ 1 \\ 1 \end{array}]$ $\[\left[\begin{array}{l}1 \\1 \\1\end{array}\right]\]$	$[\begin{array}{l} 0 \\ 1 \\ 0 \end{array}]$ $\[\left[\begin{array}{l}0 \\1 \\0\end{array}\right]\]$	64 × 179 × 179	256 × 179 × 179
4	Max pooling	3 × 3	2	0	256 × 179 × 179	256 × 89 × 89
5	ResBlock	$[\begin{array}{l} 1 \times 1 \times 128 \\ 3 \times 3 \times 128 \\ 1 \times 1 \times 512 \end{array}]$ $\[\left[\begin{array}{l}1 \times 1 \times 128 \\3 \times 3 \times 128 \\1 \times 1 \times 512\end{array}\right]\]$	$[\begin{array}{l} 1 \\ 2 \\ 1 \end{array}]$ $\[\left[\begin{array}{l}1 \\2 \\1\end{array}\right]\]$	$[\begin{array}{l} 0 \\ 1 \\ 0 \end{array}]$ $\[\left[\begin{array}{l}0 \\1 \\0\end{array}\right]\]$	256 × 89 × 89	512 × 45 × 45
6	ResBlock	$[\begin{matrix} 1 \times 1 \times 256 \\ 3 \times 3 \times 256 \\ 1 \times 1 \times 1024 \end{matrix}]$ $\[\left[\begin{array}{c}1 \times 1 \times 256 \\3 \times 3 \times 256 \\1 \times 1 \times 1024\end{array}\right]\]$	$[\begin{array}{l} 1 \\ 2 \\ 1 \end{array}]$ $\[\left[\begin{array}{l}1 \\2 \\1\end{array}\right]\]$	$[\begin{array}{l} 0 \\ 1 \\ 0 \end{array}]$ $\[\left[\begin{array}{l}0 \\1 \\0\end{array}\right]\]$	512 × 45 × 45	1024 × 23 × 23
7	ResBlock×3	$[\begin{array}{l} 1 \times 1 \times 128 \\ 3 \times 3 \times 128 \\ 1 \times 1 \times 512 \end{array}]$ $\[\left[\begin{array}{l}1 \times 1 \times 128 \\3 \times 3 \times 128 \\1 \times 1 \times 512\end{array}\right]\]$	$[\begin{array}{l} 1 \\ 1 \\ 1 \end{array}]$ $\[\left[\begin{array}{l}1 \\1 \\1\end{array}\right]\]$	$[\begin{array}{l} 0 \\ 1 \\ 0 \end{array}]$ $\[\left[\begin{array}{l}0 \\1 \\0\end{array}\right]\]$	1024 × 23 × 23	1024 × 23 × 23
8	ResBlock	$[\begin{matrix} 1 \times 1 \times 512 \\ 3 \times 3 \times 512 \\ 1 \times 1 \times 2048 \end{matrix}]$ $\[\left[\begin{array}{c}1 \times 1 \times 512 \\3 \times 3 \times 512 \\1 \times 1 \times 2048\end{array}\right]\]$	$[\begin{array}{l} 1 \\ 2 \\ 1 \end{array}]$ $\[\left[\begin{array}{l}1 \\2 \\1\end{array}\right]\]$	$[\begin{array}{l} 0 \\ 1 \\ 0 \end{array}]$ $\[\left[\begin{array}{l}0 \\1 \\0\end{array}\right]\]$	1024 × 23 × 23	2048 × 12 × 12
9	Global average pooling	12 × 12	−	−	2048 × 12 × 12	2048 × 1 × 1
10	Global average pooling	−	−	−	2048 × 1 × 1	1 × 189
11	Concat	−	−	−	1 × 189	3 × 63

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.