Ultra-low-dose PET reconstruction using generative adversarial network with feature matching and task-specific perceptual loss.

Journal: Medical physics
Published Date: Jun 17, 2019

Abstract

PURPOSE: Our goal was to use a generative adversarial network (GAN) with feature matching and task-specific perceptual loss to synthesize standard-dose amyloid Positron emission tomography (PET) images of high quality and including accurate pathological features from ultra-low-dose PET images only.

Authors

  • Jiahong Ouyang
    Stanford University, Stanford CA 94305, USA.
  • Kevin T Chen
    From the Departments of Radiology (K.T.C., F.B.d.C.M., S.S., G.Z.), Electrical Engineering (E.G., J.M.P.), and Neurology and Neurological Sciences (A.B., K.L.P., S.J.S., M.D.G., E.M.), Stanford University, 1201 Welch Rd, Stanford, CA 94305; Department of Engineering Physics, Tsinghua University, Beijing, PR China (J.X.); GE Healthcare, Menlo Park, Calif (M.K.); and Subtle Medical, Menlo Park, CA (E.G.).
  • Enhao Gong
    Department of Electrical Engineering, Stanford University, Stanford, California, USA.
  • John Pauly
    Department of Electrical Engineering, Stanford University, Stanford, CA, USA.
  • Greg Zaharchuk
    Stanford University, Stanford CA 94305, USA.

Keywords

External Resources

Popular Topics
Recent Journals