The body part regression usage pipelines.
Version 1: The body part regression model trained with basic unet with extensive datasets.
Coronal view of the BPR identification bounds from heart to pelvis.
The nii.gz format NIFTI files are required. Check where the CT images are axial reconstrcuted and in 512x512xz shape. Put CT images in
<root_path>/datasets/images
The training was performed with at least 16GB-memory GPUs.
Execute :
# cd to the bpr pipeline fodler, set root path and datasets path, run
python code/main.py --root_path <root_path> --data_dir <data_dir> --checkpoint_BPR <file path to download>Pre-trained checkpoints will be automatically downloaded.
The output BPR result list will be in:
<root_path>/txt_info/BPR_result_list.txtThe BPR doc reports will be generated in:
<root_path>/datasets/BPR_reportsA sample output report:
This is an example, not to be used for diagnostic purposes.