bm_dataset.generate_regist_pairs module

Script for generating registration pairs in two schemas

Sample run:

python generate_regist_pairs.py         -i "../output/synth_dataset/*.jpg"         -l "../output/synth_dataset/*.csv"         -csv ../output/cover.csv --mode each2all

python bm_dataset/generate_regist_pairs.py         -i "~/Medical-data/dataset_CIMA/lung-lesion_1/scale-100pc/*.png"         -l "~/Medical-data/dataset_CIMA/lung-lesion_1/scale-100pc/*.csv"         -csv ~/Medical-data/dataset_CIMA/dataset_CIMA_100pc.csv --mode each2all

Copyright (C) 2016-2019 Jiri Borovec <jiri.borovec@fel.cvut.cz>

bm_dataset.generate_regist_pairs.arg_parse_params()[source]

parse the input parameters

Return dict:parameters
bm_dataset.generate_regist_pairs.generate_pairs(path_pattern_imgs, path_pattern_lnds, mode)[source]

generate the registration pairs as reference and moving images

Parameters:
  • path_pattern_imgs (str) – path to the images and image name pattern
  • path_pattern_lnds (str) – path to the landmarks and its name pattern
  • mode (str) – one of OPTIONS_COMBINE
Returns:

DF
bm_dataset.generate_regist_pairs.main(path_pattern_images, path_pattern_landmarks, path_csv, mode='all2all')[source]

main entry point

Parameters:
  • path_pattern_images (str) – path to images
  • path_pattern_landmarks (str) – path to landmarks
  • path_csv (str) – path output cover table, add new rows if it exists
  • mode (str) – option first2all or all2all