miracle-center / ctpelvic1k Goto Github PK
View Code? Open in Web Editor NEWResources of the paper “Deep Learning to Segment Pelvic Bones: Large-scale CT Datasets and Baseline Models”.
Resources of the paper “Deep Learning to Segment Pelvic Bones: Large-scale CT Datasets and Baseline Models”.
Could you please provide the pre-trained model?
When I tried to untar dataset2, the following error occurred. Does anyone run into this issue as well?
$ tar -xzvf CTPelvic1K_dataset2_data.tar.gz.0 CTPelvic1K_dataset2_data.tar.gz.1 CTPelvic1K_datas et2_data.tar.gz.2 CTPelvic1K_dataset2_data.tar.gz.3 CTPelvic1K_dataset2_data.tar.gz.4 CTPelvic1K_dataset2_data.tar.gz.5 CTPelvic1K_dataset2_data.tar.gz.6 CTPelvic1K_dataset2_data.tar.gz.7
tar: Truncated input file (needed 109287936 bytes, only 0 available)
tar: CTPelvic1K_dataset2_data.tar.gz.1: Not found in archive
tar: CTPelvic1K_dataset2_data.tar.gz.2: Not found in archive
tar: CTPelvic1K_dataset2_data.tar.gz.3: Not found in archive
tar: CTPelvic1K_dataset2_data.tar.gz.4: Not found in archive
tar: CTPelvic1K_dataset2_data.tar.gz.5: Not found in archive
tar: CTPelvic1K_dataset2_data.tar.gz.6: Not found in archive
tar: CTPelvic1K_dataset2_data.tar.gz.7: Not found in archive
tar: Error exit delayed from previous errors.
Hi, the annotations for CLINIC-METAL looks incomplete. Out of 75 ct scans, annotations for only 13 of them are provided in the zenodo link in CTPelvic1K_dataset7_mask.tar.gz
Can you please check?
Hi,
Thank you for providing such an amazing dataset ;)
I would like to know how to compare with the provided nnUNet baseline. What's the exact dataset split? Could u provide the the data split-id correspondence so that we could compare our own method with your baseline model weights?
Cheers,
Jiancheng Yang
What should be in train_dir?
This error is reported when the original code is running.
Hi
First, thank you for the great work!
Do you maybe already have a script for the conversion of the COLONOG dataset from the multiple DICOM files to one single nii.gz file per subject that you could provide?
Kind regards
MCM-Fischer
Hi authors, thanks for the awesome work. I notice that the CLINIC and CLINIC-metal are collected in the pre-treatment and the post-treatment stages, respectively. I am wondering if there are paired data of the same patient across these two datasets? If so, could you kindly provide this information?
I tried to download sub-dataset1(ABDOMEN) and sub-dataset5(CERVIX)
but it says that "Sorry, no access to this page".
And I want to know how to download it
RuntimeError: Error(s) in loading state_dict for Generic_UNet:
size mismatch for conv_blocks_context.0.blocks.0.conv.weight: copying a param with shape torch.Size([30, 15, 3, 3]) from checkpoint, the shape in current model is torch.Size([30, 1, 3, 3]).
Is the nnunet that Pelvic1k use different from the original? Because I cannot get good result using the original. But when nnunet Pelvic1k , the segmentation is very good.
Just as the pictures shows. Both using 3D fullres structure.
It seems that original nnunet can not tell the right and left. I'm not quite familiar with the nnunet. And reading from the begining just makes me very confusing.
By the way, I'm using nnUNetTrainerV2. But I don't think it should make such a difference.
Any clue is appreciated!
I've downloaded the subdataset5 (CERVIX) , splited the test dataset out by the splits_final.pkl (fold 5).
And I downloaded the CTPelvic1K_Models.tar.gz.
I used the command_12 to run the predict, and command_16 to get the matrix.
And my mean DC is 0.84, mean HD is 18, whole DC is 0.87, whole HD is 59.
The mean DC in the paper is 0.972.
Is it normal with the 2d model and 3d cascade model, or is there something wrong with my testing method?
Thanks for this code. Are there also trained models available?
There are several data points that have a wrong physical direction:
(these are my file name but it pretty much maintain the original order)
../Data/dataset2_0003_4_325
../Data/dataset2_0460_5_325
../Data/dataset2_0480_2_324
../Data/dataset2_0616_4_325_hard
../Data/dataset2_0703_4_325
../Data/dataset2_0704_4_325
After i run one fold of 3d_fullers or another ,but this error happened to it. I have looked for a lot of solutions, but nothing seems to help the.
I tried to use different version of batchgenerators but it didn't work,
I am stuck
finished prediction, now evaluating...
Exception ignored in: <function MultiThreadedAugmenter.del at 0x7faf34d97d30>
Traceback (most recent call last):
File "/home/cui/anaconda3/envs/wyc/lib/python3.9/site-packages/batchgenerators/dataloading/multi_threaded_augmenter.py", line 269, in del
File "/home/cui/anaconda3/envs/wyc/lib/python3.9/site-packages/batchgenerators/dataloading/multi_threaded_augmenter.py", line 239, in _finish
File "/home/cui/anaconda3/envs/wyc/lib/python3.9/multiprocessing/synchronize.py", line 338, in set
File "/home/cui/anaconda3/envs/wyc/lib/python3.9/multiprocessing/synchronize.py", line 297, in notify_all
AttributeError: 'NoneType' object has no attribute 'maxsize'
Exception ignored in: <function MultiThreadedAugmenter.del at 0x7faf34d97d30>
Traceback (most recent call last):
File "/home/cui/anaconda3/envs/wyc/lib/python3.9/site-packages/batchgenerators/dataloading/multi_threaded_augmenter.py", line 269, in del
File "/home/cui/anaconda3/envs/wyc/lib/python3.9/site-packages/batchgenerators/dataloading/multi_threaded_augmenter.py", line 239, in _finish
File "/home/cui/anaconda3/envs/wyc/lib/python3.9/multiprocessing/synchronize.py", line 338, in set
File "/home/cui/anaconda3/envs/wyc/lib/python3.9/multiprocessing/synchronize.py", line 297, in notify_all
AttributeError: 'NoneType' object has no attribute 'maxsize'
Thank you for your great work in providing such a large-scale annotated dataset for the pelvis.
While working with the datasets provided, I noticed a discrepancy in the count for the COLONOG dataset. According to both the repository documentation and the associated paper, the count for this dataset is stated as 731. However, after downloading and inspecting the data, I found that there are only 714 masks.
Could you please check it for me? I'd really appreciate it.
Best
Ziyan Huang
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.