TY - JOUR
T1 - Frequency-Domain-Based Structure Losses for CycleGAN-Based Cone-Beam Computed Tomography Translation
AU - Pai, Suraj
AU - Hadzic, Ibrahim
AU - Rao, Chinmay
AU - Zhovannik, Ivan
AU - Dekker, Andre
AU - Traverso, Alberto
AU - Asteriadis, Stelios
AU - Hortal, Enrique
N1 - Funding Information:
We would like to acknowledge support from MAASTRO Clinic for the proton therapy dataset, especially Vicki Trier Taasti and Richard Canters for data curation and advisory opinions.
Publisher Copyright:
© 2023 by the authors.
PY - 2023/1
Y1 - 2023/1
N2 - Research exploring CycleGAN-based synthetic image generation has recently accelerated in the medical community due to its ability to leverage unpaired images effectively. However, a commonly established drawback of the CycleGAN, the introduction of artifacts in generated images, makes it unreliable for medical imaging use cases. In an attempt to address this, we explore the effect of structure losses on the CycleGAN and propose a generalized frequency-based loss that aims at preserving the content in the frequency domain. We apply this loss to the use-case of cone-beam computed tomography (CBCT) translation to computed tomography (CT)-like quality. Synthetic CT (sCT) images generated from our methods are compared against baseline CycleGAN along with other existing structure losses proposed in the literature. Our methods (MAE: 85.5, MSE: 20433, NMSE: 0.026, PSNR: 30.02, SSIM: 0.935) quantitatively and qualitatively improve over the baseline CycleGAN (MAE: 88.8, MSE: 24244, NMSE: 0.03, PSNR: 29.37, SSIM: 0.935) across all investigated metrics and are more robust than existing methods. Furthermore, no observable artifacts or loss in image quality were observed. Finally, we demonstrated that sCTs generated using our methods have superior performance compared to the original CBCT images on selected downstream tasks.
AB - Research exploring CycleGAN-based synthetic image generation has recently accelerated in the medical community due to its ability to leverage unpaired images effectively. However, a commonly established drawback of the CycleGAN, the introduction of artifacts in generated images, makes it unreliable for medical imaging use cases. In an attempt to address this, we explore the effect of structure losses on the CycleGAN and propose a generalized frequency-based loss that aims at preserving the content in the frequency domain. We apply this loss to the use-case of cone-beam computed tomography (CBCT) translation to computed tomography (CT)-like quality. Synthetic CT (sCT) images generated from our methods are compared against baseline CycleGAN along with other existing structure losses proposed in the literature. Our methods (MAE: 85.5, MSE: 20433, NMSE: 0.026, PSNR: 30.02, SSIM: 0.935) quantitatively and qualitatively improve over the baseline CycleGAN (MAE: 88.8, MSE: 24244, NMSE: 0.03, PSNR: 29.37, SSIM: 0.935) across all investigated metrics and are more robust than existing methods. Furthermore, no observable artifacts or loss in image quality were observed. Finally, we demonstrated that sCTs generated using our methods have superior performance compared to the original CBCT images on selected downstream tasks.
U2 - 10.3390/s23031089
DO - 10.3390/s23031089
M3 - Article
C2 - 36772129
SN - 1424-8220
VL - 23
JO - Sensors
JF - Sensors
IS - 3
M1 - 1089
ER -