Keywords:
Not applicable, Observational, Artifacts, Computer Applications-General, CT, eHealth, Computer applications, Artificial Intelligence, Artificial Intelligence and Machine Learning
Authors:
M. Patwari1, R. Gutjahr2, R. Raupach3, A. Maier1; 1Erlangen/DE, 2Nürnberg, Deutschland/DE, 3Forchheim/DE
DOI:
10.26044/ecr2020/C-03463
Results
The use of a perceptual loss function improved both feature preservation and denoising performance in 2D denoising networks (SSIM = 0.7067, PSNR = 44.5189) compared to mean squared error denoising (SSIM = 0.7065, PSNR = 44.4888).
Perceptual loss functions improved both feature preservation and denoising performance in 3D denoising networks (SSIM = 0.6944, PSNR = 45.9397) compared to mean squared error denoising (SSIM = 0.6871, PSNR = 45.6718).
Additionally, 3D networks were better at noise removal compared to 2D networks. However, the 2D networks improved feature preservation in comparison with 3D networks.
|
PSNR |
SSIM |
2D Denoising Network with MSE loss only |
44.49 ± 2.378 |
0.71 ± 0.034 |
2D Denoising Network with Perceptual Loss |
44.51 ± 2.301 |
0.71 ± 0.039 |
3D Denoising Network with MSE loss only |
45.67 ± 2.704 |
0.69 ± 0.034 |
3D Denoising Network with Perceptual Loss |
45.94 ± 2.605 |
0.69 ± 0.029 |