{"cells":[{"metadata":{},"cell_type":"markdown","source":"# The Updated HuBMAP Glomeruli FTU Segmentation Dataset\n#### by Leah Scherschel - Indiana University\n\n\nThe HuBMAP Tissue Mapping Center - Vanderbilt University (TMC-VU) team at the [BIOmolecular Multimodal Imaging Center (BIOMIC)](https://medschool.vanderbilt.edu/biomic/) has generously provided for use in this competition 10 additional PAS-stain whole slide images (WSIs), anatomical structure masks, and glomeruli segmentation masks which have never been published.\n\n## What Has Changed\n\n- The original 20 sets of images and masks (which were publicly available via the HuBMAP [Data Portal](https://portal.hubmapconsortium.org/)) have become the public train and public test datasets.\n    - 15 train\n    - 5 test\n- The 10 new, unpublished sets are the new private test dataset.\n- All 30 Glomeruli Segmentation masks (both new and original) have been thoroughly reviewed and improved. This fixed shifted masks and improved the quality of glomerulus annotations.\n- A typo in the file HuBMAP-20-dataset_information.csv for image c68fe75ea.tiff's \"width_pixels\" field has been corrected.\n- The timeline has been extended to provide competitors 2 months to work with the new data.\n   \n## What Has NOT Changed\n\n- The pipeline used to generate the WSIs.\n- The file formats.\n\nPlease refer to the prior [Dataset Details](https://www.kaggle.com/leahscherschel/dataset-details) notebook for more information on how the dataset was generated and file structure. "}],"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat":4,"nbformat_minor":4}