Code referenced in Raphael S. Daveau, et al. Deep learning based low-activity PET reconstruction of [18F]FE-PE2I and [11C]PiB in Neurodegenerative Disorders (2022), currently in the submission process.
Code in postprocessing was used for:
- image inference
- image similarity metrics computation
- clinical metrics computation
Code in segmentation was used for:
- segmentations of ROIs (used for clinical metrics)
The code used for training is found at https://github.com/CAAI/rh-torch/blob/main/rhtorch/models/attention_models_3d.py > AttentionUNet().
The original inspiration of the code comes from https://github.com/momo1689/FAN