Skip to main content

Table 5 RMSEs (HU) of reconstructed images of different noise levels

From: Low-dose CT reconstruction via L1 dictionary learning regularization using iteratively reweighted least-squares

Phantom L1-DL ADSIR GPBB SART
Shepp–Logan
 60 views, 2 million photons 10.87 31.3715 14.78 139.4
 60 views, 1 million photons 11.68 34.84 15.43 139.9
Head slice
 180 views, 2 million photons 13.46 24.81 9.582 22.86
 90 views, 2 million photons 16.83 30.54 16.83 44.49
\