search

gluucose / PCCGAN

Image2Points: A 3D Point-based Context Clusters GAN for High-Quality PET Image Reconstruction (ICASSP 2024)
12 stars 2 forks source link

About training in PCCGAN #3

Open liudelong123123 opened 2 weeks ago

liudelong123123 commented 2 weeks ago

If I need to use PCCGAN for a single mode PET image reconstruction, how do I run or call the model? Is there a code for this training?

liudelong123123 commented 2 weeks ago

Because I used your PCCGAN for training and testing, the test results are not satisfactory, and the input image pixels are normalized from 0 to 1, and the tested image pixel values are from -1 to 1200, is there something wrong?

gluucose commented 2 weeks ago

The script train.py is for sinlge model training, test.py is for multi-modal testing. Sorry for this inconsistency, I will update the code in next few days. The input images for both training and testing are all normalized to range (0,1).

liudelong123123 commented 2 weeks ago

Thank you for your correction, but it seems to me that your train.py script only trains the GAN network in the network, not your PCCGAN model, and I hope you can show the code for training PCCGans and the test code.

liudelong123123 commented 1 week ago

Can you give the training code and test code of PCCGAN?