TrainUnet

Question

TrainUnet

abinramapuram opened this issue 4 years ago · comments

Abin Abraham Oommen commented 4 years ago

I have done segmentation training with lung images and mask nodules using 2trainUnet.ipynb. But when check with test images, It just segments the lung portion, not the nodule. Did you got the nodule segmentation with the TrainUnet??

Mike Huang · Answer 1 · Tue Mar 03 2020 16:02:28 GMT+0800 (China Standard Time)

Yes, it segments the nodules. Check your labels to make sure you're training with the nodules and not the lung mask.

…

On Tue, Mar 3, 2020 at 2:46 PM Abin Abraham Oommen ***@***.***> wrote: I have done segmentation training with lung images and mask nodules using 2trainUnet.ipynb. But when check with test images, It just segments the lung portion, not the nodule. Did you got the nodule segmentation with the TrainUnet?? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#8?email_source=notifications&email_token=ACTAYDBN75LQKOSB6RNEI3TRFSRUXA5CNFSM4LADTCUKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4IR5OOIA>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACTAYDDWUTSOC5OH3BW3GMLRFSRUXANCNFSM4LADTCUA> .

Abin Abraham Oommen · Answer 2 · Tue Mar 03 2020 18:42:33 GMT+0800 (China Standard Time)

Thank you for the fast response. I have already taken the nodule masks. Number of patient samples taken is less, 30 only. Should it be the reason?? What is difference between nodulemasks and nodulemaskscircle??

Mike Huang · Answer 3 · Tue Mar 03 2020 19:32:45 GMT+0800 (China Standard Time)

See the tutorial for better understanding of the pipeline. https://www.kaggle.com/c/data-science-bowl-2017/overview/tutorial

…

On Tue, Mar 3, 2020 at 6:42 PM Abin Abraham Oommen ***@***.***> wrote: Thank you for the fast response. I have already taken the nodule masks. Number of patient samples taken is less, 30 only. Should it be the reason?? What is difference between nodulemasks and nodulemaskscircle?? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#8?email_source=notifications&email_token=ACTAYDBKHOM7J3EVAXOD5XTRFTNJVA5CNFSM4LADTCUKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOENS7FTI#issuecomment-593883853>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACTAYDB4E4ZQPKUURV6I363RFTNJVANCNFSM4LADTCUA> .