runnanchen / anatomic-landmark-detection Goto Github PK

Hi @runnanchen
Fascinating repository and article, I enjoyed reading about it and exploring your implementation.
I have successfully run your code with the same training and validation csv's, however, got quite a significant discrepancy between training and validation performance, where training is a lot higher indicating the model is overfitting. This is not too surprising given the small dataset size, but doesn't quite match the performance published, even when applying augmentation. Is there something additional that you implemented to improve performance to the level that you published? I would love to improve by validation performance so any advice would be greatly appreciated.

Hello,

Many thanks for the hard work put in the paper and code.

Would it be possible for me to have access to the drive to download the processed dataset, please?
I am unauthorized to do so currently as shown in the below image

Thank you

Hi, thank you for your contribution.

I would appreciate it if you could open issue function in the other project:
https://github.com/runnanchen/SA-LSTM-3D-Landmark-Detection

I have some questions relevent to that project.

I've been following you on this article recently. Can you open source the pddca dataset training code, the network I modified cannot achieve your results