Synthetic CT generation from CBCT using deep learning for adaptive radiotherapy in prostate cancer.

However, the slight advantage of ResU-Net over U-Net did not reach statistical significance. Although the SSIM values reported here are lower than the 0.93–0.98 range achieved by GAN-based models as described by Maspero et al., the results remain clinically acceptable given the model's simplicity and the limited dataset used ( 33 ).

also in 111,027 other papers