K-Net/contrib/MedicalSeg/configs/acdc

• Automated cardiac diagnosis
• • • Prepare dataset
  • Performance
  • • nnFormer

Automated cardiac diagnosis

ACDC involves 100 patients, with the cavity of the right ventricle, the myocardium of the left ventricle and the cavity of the left ventricle to be segmented. Each case’s labels involve left ventricle (LV), right ventricle (RV) and myocardium (MYO).The database is made available to participants through two datasets from the dedicated online evaluation website after a personal registration: i) a training dataset of 100 patients along with the corresponding manual references based on the analysis of one clinical expert; ii) a testing dataset composed of 50 new patients, without manual annotations but with the patient information given above. The raw input images are provided through the Nifti format.

Prepare dataset

To preprocess the ACDC data, you first need to download training.zip from https://acdc.creatis.insa-lyon.fr/#phase/5846c3ab6a3c7735e84b67f2

unzip training.zip
mkdir data/ACDCDataset
python tools/prepare_acdc.py training/

The dataset will be automatically automatically preprocessed. The file structure is as follows:

ACDCDataset
|--clean_data
│   ├── labelsTr
│   │   ├──patient001_frame13_0000.nii.gz
│   │   ├──patient002_frame13_0000.nii.gz
│   │   ├──patient003_frame13_0000.nii.gz
│   │   │──........
│   │   ├──patient015_frame13_0000.nii.gz
│   ├── imagesTr
│   │   ├──patient001_frame13_0000.nii.gz
│   │   ├──patient002_frame13_0000.nii.gz
│   │   ├──patient003_frame13_0000.nii.gz
│   │   │──........
│   │   ├──patient015_frame13_0000.nii.gz
├── ACDCDataset_phase
│   ├── images
│   │   ├── patient030_frame12_0000.npy
│   │   └── ...
│   ├── labels
│   │   ├── patient030_frame12_0000.npy
│   │   └── ...
│   ├── train_list.txt
│   └── val_list.txt

Then you can start the training program, such as the following command:

python train.py --config configs/acdc/nnformer_acdc_160_160_14_250k.yml --save_interval 250 --num_workers 4 --do_eval --log_iters 250 --sw_num 1 --is_save_data False --has_dataset_json False

Performance

nnFormer

Hong-Yu Zhou, Student Member, IEEE, Jiansen Guo, Yinghao Zhang, Xiaoguang Han, Lequan Yu, Liansheng Wang, Member, IEEE, and Yizhou Yu, Fellow, IEEE

Backbone	Resolution	lr	Training Iters	Dice	Links
-	14x160x160	1e-4	250000	91.78%	model| log| vdl