Conceptual
Browse Instruct Curate

Deep Learning Scatter Correction for Long-Axial-Field-of-View PET

How a convolutional U-Net can replace physics-based single-scatter simulation for correcting scattered photons in positron emission tomography, especially on long-axial-field-of-view total-body scanners where classical methods break down. The student learns how emission and attenuation sinograms are mapped to a scatter sinogram, how Monte-Carlo phantom simulations supply training data, and how scatter correction affects image quantification and lesion contrast.