something wrong in code

[nightandweather]

first of all, Thank you for sharing your code.

I'm trying to image registering in public medical dataset

https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=39879146

I want to compare how different it is from commercial software when deep learning and classical dir are applied to real rt data, not phantom.

I created a dataset with this instructions

https://github.com/junyuchen245/Preprocessed_IXI_Dataset

I converted the image resolution of (512,512,200~180) ct image to (256,256,160) cause gpu memory.(A6000*2)

and modified dataset.py, utils.py, train.py, infer.py, and config.py inside each DIR algorithm folder. modified file.pptx

Therefore, each model is producing results, but the traditional method has better results.

I'm confused, so I post a question.

Is there anything else that needs to be modified in the code?