. Author manuscript; available in PMC: 2021 Mar 1.

Published in final edited form as: IEEE Trans Med Imaging. 2020 Mar 5;39(9):2881–2892. doi: 10.1109/TMI.2020.2972547

Table 1.

Artefact removal performance on hold-out simulated and real-world artefacts presented as median [1st quartile, 3rd quartile]. Pairwise statistically significant according to paired sample Wilcoxon Tests (Bonferroni corrected) Improvements are Indicated in bold.

	Simulated
	Uncorrected	Corrected
MAE	0.00434 [0.00323, 0.00599]	0.00343 [0.00271, 0.00452]
SSIM	0.96835 [0.94835, 0.98338]	0.98989 [0.98193, 0.99377]

	Real-world
	Uncorrected	Corrected
MAE	0.00302 [0.00231, 0.00340]	0.00297 [0.00230, 0.00338]
SSIM	0.98566 [0.98001, 0.98886]	0.98604 [0.98105, 0.98960]