In the article, the authors write that "The test images are divided into overlapping patches for testing in the real implementation." I question if the "overlap" operation for testing an image increases the total amount of information, i.e. if the real sampling ratio is greater than the setting? Thanks for your reply.
In the article, the authors write that "The test images are divided into overlapping patches for testing in the real implementation." I question if the "overlap" operation for testing an image increases the total amount of information, i.e. if the real sampling ratio is greater than the setting? Thanks for your reply.