Skip to main content
. 2019 Sep 13;14(9):e0222406. doi: 10.1371/journal.pone.0222406

Fig 9. Comparisons of the artifact indices (AIs) determined for in-focus plane images procured via the denoising convolutional neural network metal artifact reduction hybrid reconstruction (DnCNN-MARHR) algorithm and the traditional reconstruction algorithms with metal artifact reduction (MAR) processing.

Fig 9

Metal artifacts originated from the AIs of 10 selected metal artifact areas (features) and one background area are presented in the in-focus plane. The results are mean ± standard error.