Fig. 9From: Clinical and phantom validation of a deep learning based denoising algorithm for F-18-FDG PET images from lower detection counting in comparison with the standard acquisitionScatter plot of the variation of Contrast to Noise Ratio between the PET100 and the PET50 ± SP images, as a function of the PET100 SUVmax, for the 60 lesions selected by the physicians. The CNR restoration is less dependent to lesion SUVmax with SubtlePETBack to article page