lucasrla / wsi-preprocessing

Simple library for preprocessing histopathological whole-slide images (WSI) into tiles (a.k.a. patches) towards deep learning

wsi-preprocessing

Simple library for preprocessing histopathological whole-slide images (WSI) towards deep learning.

This project is a series of simple scripts that wraps pyvips and builds upon lucasrla/wsi-tile-cleanup.

Features:

• Find the whole-slide image (WSI) resolution (in microns per pixel, MPP) via OpenSlide
• Calculate the Otsu threshold for whole-slide image (WSI) via lucasrla/wsi-tile-cleanup
• Set the target MPP (say, 0.5) and the tile dimensions (say, 224px) and let libvips do its (fast!) magic to create tiles from whole-slide images (WSI)
• Compute masks on tiles for filtering out unneeded tiles via lucasrla/wsi-tile-cleanup

For an example of an actual pipeline (built with SoS Workflow) that uses wsi_preprocessing, see lucasrla/wsi-preprocessing-sos-workflow.

Installation

Conda

conda create --name YOUR_ENV_NAME --channel conda-forge python=3.6 libvips pyvips numpy
conda activate YOUR_ENV_NAME

# note: `python3.6 -m pip` is just to make sure we are using pip from python=3.6
python3.6 -m pip install git+https://github.com/lucasrla/wsi-preprocessing.git

pip or poetry

# first of all, install libvips 
# https://libvips.github.io/libvips/install.html
# (tip: have it installed with openslide support)

# next, create a new virtualenv and activate it using your tool of choice
# (e.g., pyenv, virtualenv, etc)

# then, depending on your dependency manager, run either:
poetry add git+https://github.com/lucasrla/wsi-preprocessing.git
# or
pip install git+https://github.com/lucasrla/wsi-preprocessing.git

Usage

If you don't have a whole-slide image available in your local machine, you need to download one first. I suggest you this one: http://openslide.cs.cmu.edu/download/openslide-testdata/Aperio/CMU-1.svs (170 MB). It is hosted by the folks that created the great OpenSlide at CMU.

# example.py

import wsi_preprocessing as pp

# if, for instance, CMU-1.svs is in your current directory ("."):
slides = pp.list_slides(".")
pp.save_slides_mpp_otsu(slides, "slides_mpp_otsu.csv")

# this may take a few minutes (depending on your local machine, of course)
pp.run_tiling("slides_mpp_otsu.csv", "tiles.csv")

pp.calculate_filters("slides_mpp_otsu.csv", "", "tiles_filters.csv")

python3.6 example.py

Voilà! After running example.py, you will have the defaults:

• png tiles (at 0.5 mpp) that have 224x224 pixels generated from your whole-slide image
• csv files with metadata that will enable you filter out most of the gibberish

Credits and Acknowledgments

Just like wsi_tile_cleanup, please note that wsi_preprocessing is just a very thin wrapper around libvips, pyvips and numpy. They are the ones doing the heavy lifting (and doing it amazingly well).

Besides them, I should also mention:

• OpenSlide: makes it possible to do digital pathology on an open source environment. They also host several test files that you easily can peruse online.

License

wsi-preprocessing is Free Software distributed under the GNU General Public License v3.0.

Dependencies have their own licenses, check them out.