Skip to main content
. Author manuscript; available in PMC: 2021 Feb 26.
Published in final edited form as: IEEE Trans Comput Imaging. 2020 Apr 27;6:843–856. doi: 10.1109/TCI.2020.2990299

Table III. Scaling properties of discussed algorithms in terms of memory footprint and operator evaluations. Here, n is the image size, d is the image dimension (usually d = 2,3), and N refers to number of unrolled iterations in learned iterative schemes.

Memory OPERATOR EVAL.
FBP O(nd) 1
U-Net denoiser O(nd) 1
LGS O(Nnd) O(N)
Multi-scale O(nd) O(1)