/brainPETNR

Code used in relation to the neurodegenerative disorder study with [11C]PiB and [18F]FE-PE2I tracers (2022). Image preparation, model inference and metrics extraction.

Primary LanguagePython

Brain PET Noise Reduction (brainPETNR)

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)

Model

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