Skip to main content
. 2019 Mar 8;9:3926. doi: 10.1038/s41598-019-40554-1

Figure 2.

Figure 2

Schematic of the training process for deep-learning based pixel super-resolution. An outline of the data required to generate the network input and ground truth images is shown, together with an overview of how both the traditional super-resolution is performed and how the deep learning super-resolution network is trained.