hilab-git / scpm-net Goto Github PK

Hi, I am very interested in your work!
I tried your code on the previous structure DeepLung released on Github. I found the output of the SCPM-Net is C1, R1, off1, C2, R2, and off2. But how do they refuse together? I did not find it out.
And the Center Point matching strategy is essential to select positive and negative samples so that the loss in the training process can be calculated. So could you release your code of the Top-K center point matching strategy? I did not figure out the matching strategy.
Thanks a lot.
Look forward to hearing from you~!

Hi👋, I am very interested in your work！I would like to ask if it is possible to release the whole training code publicly. If there is no hope for the release of the entire code, can I ask for a private copy of the code and study it.
Looking forward to your prompt reply!😊

Hi，
Thanks for your awesome work. I am confused about the shape of output. The shape of input is (96, 96, 96), but that of the output is (48, 48, 48) in SCPMNet.py. The shapes of input and output should be the same?

Hi, I'm very interested in your work. Could you possibly make your entire training code public?

Hi，
Thanks for your awesome work. I am confused about the ground truth.Could you tell me how to sample one patch from the luna16 dataset, and map the ground truth of one patch with the nodule location and size annotation, which contain one more nodule?

I am very interested in your paper published in MICCAI 2020. When will you open the code?

Thanks,

I think p_j should be 1-p when it's negative sample and p in positive sample, p is the probability of being a nodule. @Luoxd1996

I am very interest in nodule detection in CT.
Thanks for this work.
I would like to ask if it is possible to release the whole training code publicly.
If there is no hope for the release of the entire code, can I ask for a private copy of the code and study it.
I hope to reply. Thank you

大神你好可以开源你的全部代码吗