Skip to main content
. 2018 Dec 28;20(2):295–303. doi: 10.3348/kjr.2018.0249

Table 3. RMSE Values of Subtracted CT Volumes in Test 2.

Conversion Pairs (Input, Target Image) Original Converted Differences (%)
From smooth to sharp kernel
 B10f, B30f 15.02 ± 0.87 5.17 ± 0.45 −65.6 ± 2.3
 B10f, B50f 53.50 ± 5.08 28.84 ± 4.52 −46.4 ± 4.0
 B10f, B70f 109.50 ± 14.07 74.97 ± 13.46 −31.9 ± 4.0
 B30f, B50f 40.68 ± 4.65 15.72 ± 2.15 −61.4 ± 1.5
 B30f, B70f 98.50 ± 13.94 45.49 ± 7.47 −53.9 ± 1.5
 B50f, B70f 60.63 ± 9.56 18.77 ± 2.70 −69.0 ± 0.7
From sharp to smooth kernel
 B30f, B10f 15.02 ± 0.87 3.44 ± 0.18 −77.1 ± 0.8
 B50f, B10f 53.50 ± 5.08 10.85 ± 1.16 −79.7 ± 1.1
 B50f, B30f 40.68 ± 4.65 8.81 ± 1.10 −78.3 ± 1.3
 B70f, B10f 109.50 ± 14.07 19.92 ± 2.15 −81.7 ± 1.3
 B70f, B30f 98.50 ± 13.94 17.55 ± 1.56 −82.0 ± 1.5
 B70f, B50f 60.63 ± 9.56 13.63 ± 1.63 −77.3 ± 1.3

Data are presented as mean ± SD (HU). Original: RMSE by subtracting between input and target images. Converted: RMSE by subtracting converted images from target images. Differences: 100 x (converted RMSE - original RMSE) / original RMSE