vandermode / tfpnp Goto Github PK

RL need reword which is PSNR. Thus, how can we deal with the data which do not have label?

Thanks for your contribution, it is wonderful.I wanted to use your code as a comparison algorithm for CT reconstruction, but I found that only PSNR but no SSIM was calculated in the code.I see PSNR is calculated by "eval_psnr = self.evaluator.eval(self.actor, step)", but I don't know how to calculate SSIM by using "self.actor" and "step".I would appreciate it if you could tell me.

Thank you for your fascinating paper.
When I want to train CT dataset, where the code from script.sh, there is a mistake. "AttributeError: module 'tfpnp.utils.transforms' has no attribute 'RadonGenerator'".
1 I woder, should I set enable_CT = True(line 445, in transforms.py)？
2 But when I set enable_CT = True(line 445, in transforms.py), there is a new mistake. "NameError: name 'Radon' is not defined". (PS: have downloaded torch_radon in url:https://github.com/kshkss/torch-radon)
Looking forward to your reply!

Thank you for your fascinating paper.
When I want to run the 'csmri' testing code in the new version, where the code from script.sh, there is a mistake. "TypeError: 'CSMRIEnv' object is not iterable".
Looking forward to your reply!

如题，我在实验中发现最大时间步为1效果好像更好

Thanks for your contribution, it is wonderful.I wanted to use your code as a comparison algorithm for CT reconstruction, so I want to calculate the number of parameters in TFPnP.I tried to calculate it by trainer.parameters() but failed.So I want to know calculate the number of parameters in TFPnP.I would appreciate it if you could tell me.

Thanks for your contribution, it is wonderful.I wanted to use your code for the reconstruction of sparse-view CT with metal artifacts on deeplesion dataset, but the results are not good.At the 800th step of training, psnr is still lower than 20dB.And the psnr increases by no more than 1dB after many steps of training, and sometimes decreases.I want to know if I should adjust the hyperparameter or learning rate.I would appreciate it if you could tell me.

As the paper said “For the CS-MRI application, a single policy network is trained to handle multiple sampling ratios (with x2/x4/x8 acceleration) and noise levels (5/10/15), simultaneously. Similarly, one policy network is learned for phase retrieval under different settings. “, does the motioned training datasets of applications (MRI, PR, CT, SPI) are all PASCAL VOC dataset? Because there is a domain shift between the training datasets and the testing dataset in (MRI, CT, SPI).