Skip to main content
. 2024 Feb 16;14:3917. doi: 10.1038/s41598-024-54649-x

Table 3.

SSIM and PSNR values in each dispersed projection angle with artifact reduction by deep learning models. These values were calculated based on the original CT image. *p-values of < 0.005 were deemed significant for differences from the results of corresponding results of artifacts image.

Interval of projection data acquisition angle
10°
SSIM
 With artifacts 0.87 ± 0.02 0.58 ± 0.02 0.29 ± 0.02 0.17 ± 0.02
 AE 0.40 ± 0.03* 0.38 ± 0.02* 0.37 ± 0.02* 0.31 ± 0.03*
 U-Net 0.41 ± 0.03* 0.41 ± 0.03* 0.40 ± 0.03* 0.31 ± 0.02*
 CGAN 0.94 ± 0.02* 0.90 ± 0.03* 0.87 ± 0.02* 0.79 ± 0.02*
PSNR
 With artifacts 33.6 ± 0.3 26.8 ± 0.3 19.7 ± 0.4 16.0 ± 0.5
 AE 24.8 ± 0.1* 24.5 ± 0.1* 23.3 ± 0.2* 20.4 ± 0.4*
 U-Net 24.9 ± 0.1* 24.8 ± 0.1* 24.0 ± 0.1* 21.1 ± 0.3*
 CGAN 36.7 ± 0.7* 33.3 ± 0.4* 29.3 ± 0.4* 23.7 ± 0.6*

*p < 0.005, showing a significant difference from the artifact conditions.