- pydicom
- nibabel
- pandas
- numpy
dcm2niix_HsinChu.py- This script has been abandoned.
dcm2niix_Taipei.py-
Do not use this script for decompression.
-
The dictionary should be like:
/media/shard/data/ct_scan/raw/Taipei_2897/CGMH <-- That is for argument `--data_root` ├── 0001 <-- here are the patient ├── 0003 └── ...... /media/shard/data/ct_scan/raw/Taipei_2897/ISP_annotation <-- That one should pass into `--isp_root` ├── 0001 # Almost like 502-CT ├── 0003 └── ......- PS: To ignore parts that have already been processed, the
--ignore_pathshould point to the--meta_dir.
- PS: To ignore parts that have already been processed, the
-
The progress
- A Skip ISP status indicates that the ISP folder does not contain any tissue labels.
-
python store_each_plq.py ...
python combine_multi_plque.py ... - Choosing a Script Based on CT Imaging Source (Taipei or Hsinchu)
- For Taipei:
Specify the
--meta_pathargument. Its value should be the same as the--metaargument used indcm2niix_Taipei.py.- Example:
python text_match/maek_taipei_image_table.py \ --meta_path <path/to/your/meta/data>
- Example:
- For Hsinchu:
Specify the --hsinchu_root argument. Its value should be the same as the dst_root argument used in dcm2niix_Taipei.py.
- Example:
python text_match/make_hsinchu_image_table.py \ --hsinchu_root <path/to/your/destination/root>
- Example:
- For Taipei:
Specify the
- Merging Matched Image Cases and Reports
To change the matching conditions, please modify the
get_match_dfmethod.
- Example:
python text_match/merge_image_report.py \ --image_table_path <Path to the output from make_taipei_image_table.py or make_hsinchu_image_table.py> \ --report_label_path <Path to the desired report storage file> \ --mode [taipei / hsinchu]
- The dcm2niix program is available for download at dcm2niix
- Cardiac Structure ID to Name
- Right Atrium
- Right Ventricle
- Left Atrium
- Left Ventricle
- MyocardiumLV
- Aorta
- Coronaries8
- Fat
- Bypass
- Plaque