Deep learning-based imaging in radio interferometry

Free Access

Table 1

Overview of the used architecture.

Stage	Layer	Channel in	Channel out	Kernel size	Stride	Padding
Pre block	Convolution PReLU	2 64	64 64	(9 × 9)	1	4

Residual block (×8)	Convolution Batch Norm PReLU Convolution Batch Norm Elementwise Sum	64 64 64 64 64 64	64 64 64 64 64 64	(3 × 3) – – (3 × 3) – –	1 – –1 – –	1 – –1 – –

Post Block	Convolution Batch norm Elementwise sum	64 64 64	64 64 64	(3 × 3)	1	1

Final block	Convolution	64	2	(9 × 9)	1	4

Notes. For each layer, the number of input and output channels is given. In the case of convolutional layers, the kernel size, stride, and padding are specified.

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.