Group equivariant pyramid network for respiratory motion correction on PET image

Key Points

• Respiratory motion correction significantly improves PET image quality, leading to more accurate diagnoses.
• The pyramid network models effectively reduce motion artifacts in PET scans, enhancing reliability in clinical settings.
• Using a group equivariant framework, the model demonstrates advanced capabilities in maintaining spatial information during image correction.
• Highlighting the need for advanced imaging techniques, the findings support future investigations for broader clinical applications.