Code for extracting ROI from annotated structures in whole slide images (WSIs) using Aperio Imagescope
Place 'extract_xml_region.py' in a directory containing the .svs WSI and .xml annotation files before running. This code contains options for extraction at the top:
- save_dir ==> output directory name - created
- size_thresh ==> size threshold in pixels - excludes small objects
- final_image_size ==> size of the output ROI patches
- white_background ==> mask the structure - set the background to white (set to False for whole ROI)
- extract_one_region ==> include only one annotation per ROI (white_background must be set to True)
Code that can be trained to segment WSIs and produce XML annotations is avalable here
This code was created by Brendon Lutnick