search

shaoyanpan / Synthetic-CT-generation-from-MRI-using-3D-transformer-based-denoising-diffusion-model

This is the repository for the paper published in Medical Physics: "Synthetic CT generation from MRI using 3D transformer-based denoising diffusion model".
https://aapm.onlinelibrary.wiley.com/doi/abs/10.1002/mp.16847
MIT License
30 stars 7 forks source link

Use your model to genereate PET by MRI #6

Open MaoFuyou opened 4 months ago

MaoFuyou commented 4 months ago

I use your model ,train paired MRI-PET nii data, but get an extremely bad output! I want to know why and how to deal with it. training_log.csv

MaoFuyou commented 4 months ago

So, In total, I have 220 paired MRI-PET data, What advice do you have for spliting trainning patient and testing patient?

shaoyanpan commented 4 months ago

You have to make 70:10:20 for training, testing and validation, in terms of writing paper. Even worse in clinical journals, you may need to do five-fold cross validation otherwise no one trust you. Unless you are very good at clinical application, you can manually select some few very clinical meaningfull patients for testing then you can use all the rest for training......

MaoFuyou commented 4 months ago

Ok, Thank your for detailed explaination.