b4edda67f0f57e218a8e766927e3e5c5-AuthorFeedback.pdf

We thank the reviewers for their constructive comments. We provide our answers below. Presentation: if the paper is accepted, we will improve our presentation, notably by better explaining "the big picture" R1: "The authors consider both gradient l1 an l2 type regularization, as well as "standard" l1 and l2 methods [..]". "PSNR values by themselves don't really provide an overly compelling justification". Finally, we also plan to include SSIM scores.