This repo contains unofficial fixes for the Kaggle Data Science Bowl 2018 stage 1 train dataset. The dataset was created by the Carpenter lab at the Broad Institute of Harvard and MIT for 2018 Data Science Bowl and released under Creative Commons license 0 (CC0 public domain): https://www.kaggle.com/c/data-science-bowl-2018/discussion/47864

Fixes are released under the same license: Creative Commons license 0 (CC0 public domain). All fixes are in master, this is a drop-in replacement for the original dataset. You don't need to use mask_tools.py in order to use the dataset - these tools were used for creating the annotations. Pull requests welcome!