Segmenatation is based on three datasets:
Note, datasets are in .gitignore
use the following link to get data: datasets.
After that, add datasets to root folder RetinaVesselSeg_tf
- clone main branch
- Download the data to main
- run command
python .\data_processing.py -s
in main folder. It createsdata_paths.csv
file with structureNAME, DATASET_NAME, PATH_TO_ORIGINAL_IMAGE, MASK
Alternative name of csv,e.g., data.csv, is also possible bypython .\data_processing.py -c 'data.csv'
ToDo: images, masks to tfrecods