DTM

DTM: Diffusion Transformer Model Guided by Compact Prior in Low-dose PET Reconstruction
The Code is created based on the method described in the following paper:
Diffusion Transformer Model Guided by Compact Prior in Low-dose PET Reconstruction
B. Huang，X. Liu，L. Fang，Q. Liu, B. Li
Phys Med Biol. https://iopscience.iop.org/article/10.1088/1361-6560/adac25

Optional parameters:

weight: Weight for forward loss.
epoch: Specifies number of iterations.

The training pipeline of DTM

Two visualization pipeline of DTM

The results of PET images

Training

1. To pretrain DTM_S1, run

sh trainS1.sh

4. To train DTM_S2, run

#set the 'pretrain_network_g' and 'pretrain_network_S1' in ./options/train_DTMS2.yml to be the path of DTM_S1's pre-trained model

sh trainS2.sh

Note: The above training script uses 8 GPUs by default.

Testing

• Testing

# modify the dataset path in ./options/test_DTMS2.yml

sh test.sh 

Other Related Projects

• ALL-PET: A Low-resource and Low-shot PET Foundation Model in Projection Domain [Paper] [Code]

• RED: Residual Estimation Diffusion for Low-Dose PET Sinogram Reconstruction [Paper] [Code]

• Double-Constraint Diffusion Model with Nuclear Regularization for Ultra-low-dose PET Reconstruction [Paper] [Code]

• Diffusion Transformer Meets Random Masks: An Advanced PET Reconstruction Framework [Paper] [Code]

• Raysolution_PET_Data [Data]

• Synthetic CT Generation via Variant Invertible Network for Brain PET Attenuation Correction [Paper] [Code]

• Spatial-Temporal Guided Diffusion Transformer Probabilistic Model for Delayed Scan PET Image Prediction [Paper] [Code]

• PET Tracer Separation using Conditional Diffusion Transformer with Multi-latent Space Learning [Paper]

• Temporal Image Sequence Separation in Dual-Tracer Dynamic PET With an Invertible Network [Paper] [Code]

• A Prior-Guided Joint Diffusion Model in Projection Domain for PET Tracer Conversion [Paper] [Code]

• Positron Emission Tomography Tracer Conversion via Variable Augmented Invertible Network [Paper]