slideio

SlideIO Python Wrapper

This repository provides a Python interface to the SlideIO library, a high-performance C++ image-reading framework designed for medical and biological slides. With SlideIO, you can read whole-slide images, extract specific regions, and scale large slides efficiently. The Python wrapper integrates these capabilities into a convenient API, returning image data as NumPy arrays.

License: BSD-3-Clause.

---

References

• SlideIO GitHub Repository
• SlideIO Tutorial
• SlideIO Website
• SlideIO Python Documentation
• Using SlideIO Python Library for Image Analysis in Digital Pathology
• SlideIO: A New Python Library for Reading Medical Images

---

Overview

SlideIO is available as:

• A C++ library for efficient reading of whole-slide images, with support for multi-dimensional data (2D, 3D, and time series).
• A Python module that leverages the same C++ codebase, providing a Pythonic interface for reading images into NumPy arrays.

Key features include:

• Whole-slide reading: Load entire slides or specific regions of interest.
• Efficient scaling: Rapidly generate scaled images from large slides using internal zoom pyramids.
• Format flexibility: Handle diverse medical and biological imaging formats through an extensible driver system.

Supported Image Formats

The following table lists the currently implemented drivers and their corresponding file formats:

Driver	File Format	Extensions	Developer	Scanners
SVS	Aperio SVS	*.svs	Leica Microsystems	Aperio GT 450 and Aperio GT 450 DX
AFI	Aperio AFI (Fluorescent)	*.afi	Leica Microsystems
SCN	Leica SCN	*.scn	Leica Microsystems	Leica SCN400
CZI	Zeiss CZI	*.czi	Zeiss Microscopy	ZEISS Axioscan 7
ZVI	Zeiss ZVI	*.zvi	Zeiss Microscopy
DCM	DICOM	*.dcm / no extension	-	-
NDPI	Hamamatsu NDPI	*.ndpi	Hamamatsu
VSI	Olympus VSI	*.vsi	-	-
QPTIFF	PerkinElmer Vectra QPTIFF	*.qptiff	Akoya Biosciences	PerkinElmer Vectra
GDAL	Common image formats (JPEG, PNG, TIFF, etc.)	*.jpeg, *.jpg, *.tif, *.tiff, *.png	-	-

To learn more about the library and additional features, visit the SlideIO Website.

---

Building on Linux with the Manylinux Docker Container

Below are instructions for building the Python wheels within a manylinux environment.

1. Clone Repositories

   git clone https://github.com/Booritas/slideio-python.git
   git clone https://github.com/Booritas/conan-center-index.git
   

2. Run the Docker Container
   Map the parent directory containing these repositories into the container:

   docker run --name slideio -it \
     -v /path-to-slideio-python-parent-directory:/slideio \
     booritas/slideio-manylinux_2_28_x86_64:2.7.0 bash
   

3. Set Environment Variables
   Within the container, point to the cloned repositories:

   export SLIDEIO_HOME=/slideio/slideio-python
   export CONAN_INDEX_HOME=/slideio/conan-center-index
   

4. Build Custom Conan Packages

   cd /slideio/slideio-python
   ./conan.sh
   

5. Install Conan Dependencies

   cd /slideio/slideio-python
   python3 ./install.py -a conan -c release
   

6. Build Python Wheels

   cd /slideio/slideio-python
   ./build-wheels-manylinux.sh
   

7. Locate the Wheel Packages
   You can find the resulting wheel files in the wheelhouse subdirectory of the slideio-python repository.

---

Contributing

Please submit pull requests or open issues if you have any suggestions or find any bugs. We welcome contributions from the community to improve the library and its Python bindings.

---

Enjoy using SlideIO for your medical and digital pathology workflows!