Fig. 3From: Artifact-free fat-water separation in Dixon MRI using deep learningPredicted values were performed on a participant in the test set, where the model was trained on all 800Â participants in the training set. Absolute differences are displayed in the original image intensities. Performance metrics calculated on the 3DÂ volume are: \(\text {SSIM}_W=0.953\), \(\text {SSIM}_F=0.970\), \(\text {PSNR}_W=29.80\,\text {dB}\) and \(\text {PSNR}_F=29.85\,\text {dB}\)Back to article page