Instructions on install/run your code?

[tianqig]

Hi

Thanks for your excellent works and sharing your code! I am very interested in your paper and would like to try it myself, wonder if you could add more details as to versions of all related modules such as tensorflow, opencv, etc., examples of steps to segment a few MRI/CT images, display segmentation results, etc.

It would be great to show with Jupyter notebooks. Also I wonder if your code can be run on CPU only?

Thanks
David

[xy0806]

Dear David, thanks for you interest in our code. For more details, you can refer to our previous publications: (a) 3D U-net with Multi-level Deep Supervision: Fully Automatic Segmentation of Proximal Femur in 3D MR Images (b) 3D FractalNet: Dense Volumetric Segmentation for Cardiovascular MRI Volumes and this code related paper which will online in December (c) Hybrid Loss Guided Convolutional Networks for Whole Heart Parsing <https://www.researchgate.net/publication/319702011_Hybrid_Loss_Guided_Convolutional_Networks_for_Whole_Heart_Parsing> For the instruction of code usage, I think the code itself is self-explanatory with the .ini file, just run the main.py. I use CV2, tf 1.0. I didn't test CPU version, since it's too slow. If you need more help, you can contact me via email: yangxinknow@gmail.com, or WeChat: 352268699. best, xin

…

On 1 November 2017 at 13:38, David Tian ***@***.***> wrote: Hi Thanks for your excellent works and sharing your code! I am very interested in your paper and would like to try it myself, wonder if you could add more details as to versions of all related modules such as tensorflow, opencv, etc., examples of steps to segment a few MRI/CT images, display segmentation results, etc. It would be great to show with Jupyter notebooks. Also I wonder if your code can be run on CPU only? Thanks David — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#2>, or mute the thread <https://github.com/notifications/unsubscribe-auth/APZy15hbzqwtupZ2v_QyJiAV_DOoyuAMks5syAPhgaJpZM4QNxdO> .

[tianqig]

Dear Xin
Thanks for your quick reply and detailed info, very helpful.
I will contact you if need more help.

Best

David

[fishfishson]

Dear Xin,
I run your code and find something wrong with training. The GPU memory of the model is only around 170MB each and the training result is strange, only background and LV has non-zero dice. Do you test your code on this repo?

Best
Zhiyuan

[xy0806]

Dear Zhiyuan, the model has been tested. maybe you are using the CPU rather than GPU. if everything is correct, you will see the DICE increases and then decreases and finally gradually increases to convergence. best, xin

…

On Wed, 17 Jul 2019 at 16:07, Zhiyuan ***@***.***> wrote: Dear Xin, I run your code and find something wrong with training. The GPU memory of the model is only around 170MB each and the training result is strange, only background and LV has non-zero dice. Do you test your code on this repo? Best Zhiyuan — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2?email_source=notifications&email_token=AD3HFVZBQB6HYAE2SJIZ5E3P73HKPA5CNFSM4EBXC5HKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2DMP3Y#issuecomment-512149487>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AD3HFV3PMBJMMVWGYJXOULDP73HKPANCNFSM4EBXC5HA> .