TheLion-ai / UMIE_datasets

Creative Commons Attribution 4.0 International
33 stars 1 forks source link

[CMMD] Create a pipeline for CMMD dataset #91

Closed KacperRogala closed 1 month ago

KacperRogala commented 2 months ago

Issue: #80

Description: Creates a preprocessing pipeline for images stored in CMMD dataset. The purpose is to add the ability to preprocess another medical image dataset to the repository. The dataset is CMMD (The Chinese Mammography Database) containing 5,202 MG (Mammography) breast images.

Additionally the file steps/convert_dcm2png.py is changed. Previously, for this step to work correctly, the masks_path argument had to be provided in the config/runner_config.py file, even if the dataset did not have masks. Now the masks_path argument is optional.

andrekomor commented 1 month ago

Make sure testing step is passing before merging

KacperRogala commented 1 month ago

Make sure testing step is passing before merging

Honestly, I don't know why the test is failing. Based on the fail message, I understand that it can't find the CMMD_clinicaldata_revision.xlsx file. But it should be in the test files, at least it is on the drive.