Fig. 7From: Combining deep learning with a kinetic model to predict dynamic PET images and generate parametric imagesA comparison of the image quality of the SUV images produced with and without a kinetic model, where a real 1-h static PET image was used as a reference. The suffix “_h” means that the higher this metric is, the better the image quality. In contrast, the suffix “_l” means that this metric is a distortion index, so the lower this metric is, the better the image quality. The assessment of the radar image on the right shows that the larger the footprint is, the better the image qualityBack to article page