Medisys

Yonsei Medical Imaging Systems Lab

A Performance Comparison of CNN-based Image Denoising Methods: The Effect of Loss Functions on Low-dose CT Images

B. Kim, M. Han, H. Shim, and J. Baek | Med Phys, 2019

Figure 5: The XCAT test image denoised in the image domain for three different dose levels. (a) LDCT, (b) CNN-MSE, (c) CNN-MAE, (d) CNN-VGG, (e) CNN-VGGMSE, (f) CNN-VGGMAE, (g) WGAN-VGGMSE, (h) WGAN-VGGMAE, (i) TV-IR, (j) NDCT. The red dotted circles indicate four ROIs that contain the inserted signals. The display window is (-160, 240) in HU units.
Figure 7: The XCAT test image denoised in the sinogram domain for three different dose levels. (a) LDCT, (b) CNN-MSE, (c) CNN-MAE, (d) CNN-VGG, (e) CNN-VGGMSE, (f) CNN-VGGMAE, (g) WGAN-VGGMSE, (h) WGAN-VGGMAE, (i) TV-IR, (j) NDCT. The red dotted circles indicate four ROIs that contain the inserted signals. The display window is (-160, 240) in HU units.
Figure 9: Radial NPS curves from different denoising methods for three different dose levels. The CNN-based denoising methods were applied to (a-c) reconstructed images and (d-f) sinogram images. The dose levels of input images were (a) and (d) 25%, (b) and (e) 50%, and (c) and (f) 75% of the normal dose level.
Figure 10: Radial MTF of CNN-based denoising methods for 60 HU and 160 HU signal contrast levels and 25%, 50%, and 75% dose levels. (a) CNN-VGG and (b) CNN-VGGMSE performed on the image domain and (c) CNN-VGG and (d) CNN-MAE performed on the sinogram domain.

A performance comparison of anthropomorphic model observers for breast cone beam CT images: A single-slice and multislice study

M. Han and J. Baek | Med Phys, 2019

FIG. 2. Comparison of the Pc of single-slice images between the human observers and (a) D-DOG ssCHO-ci (b) Gabor ssCHO-ci, (c) D-DOG ssCHO-di, (d) Gabor ssCHO-di, (e) ssNPWE4-di, and (f) ssNPWE7-di.
FIG. 5. Comparison of the Pc of multislice images between the human observers and (a) D-DOG msCHO-ci, (b) Gabor msCHO-ci, (c) D-DOG msCHO-di, (d) Gabor msCHO-di, (e) msNPWE4-di, and (f) msNPWE7-di.

Multi-pass approach to reduce cone-beam artifacts in a circular orbit cone-beam CT system

C. Han and J. Baek | Optics Express, 2019

Fig. 12. Coronal (upper) and sagittal (lower) aspects of the XCAT phantom: (a) the reference and (b) FDK images, and corrected images using (c) the two-pass algorithm and the proposed algorithm after (d) the 1st iteration and (e) the 4th iteration and (f) the TV-IR algorithm. Units are in cm^-1 and the display window was [0.1 0.3].
Fig. 15. Coronal (upper) and sagittal (lower) aspects of the pediatric phantom: (a) FDK images and corrected images using (b) the two-pass algorithm and the proposed algorithm after (c) the 1st iteration and (d) the 5th iteration and (f) the TV-IR algorithm. Units are in cm^-1 and the display window was [0.1 0.3].