HistoClean Versions Save

HistoClean is a tool for the preprocessing and augmentation of images used in deep learning models. This easy to use application brings together the most popular image processing packages from across the python universe, meaning no more looking at documentation! HistoClean provides real time feedback to augmentations and preprocessing options. This allows users to evaluate their steps before implementation.

v0.1.3

2 years ago

A few bug fixes added:

  • Fixed a bug where the app would not accept images wit capitalized extensions.
  • Fixed a bug in the crop feature where the crop location was incorrect
  • Fixed a bug in the normalization module where the progress bar would not update.
  • Fixed a bug in the augmentation module where single channel images would not register.

v0.1.2

2 years ago

Downloads

Windows: HistoClean 0.1.2.zip

Changes:

  • New Streamlined Interface
  • Added OpenSlide Support
  • Image patching now supports .tif and .ndpi file formats
  • Whitespace filter can now create binary masks
  • Whitespace filter now supports .tif format
  • Fixed several bugs around the quick balance module
  • Improved error handling across the application
  • Added Crop and resize options to image pre-processing
  • Fixed Image Augmentation bug that created a new folder for each image for certain augmentations

Please note that newest additions have slightly increased startup time by a few seconds. We are looking to improve this in future updates

v0.1.1

2 years ago

First official update to HistoClean

Downloads

Windows: HistoClean 0.1.1.zip

Changes:

  • Improved startup times and splash screen.
  • Reduced program size.
  • Reinhard Normalisation option added.
  • Added augmentation option "Add to existing image set".
  • Various bug and performance fixes.

It is recommended to make a copy of your images before applying HistoClean to avoid accidental data loss.

0.1.0

3 years ago

This is the initial public pre-release for HistoClean. This version is for testing and is the version used in the paper "HistoClean: open-source software for histological image pre-processing and augmentation to improve development of robust convolutional neural networks "

In this initial release, user can:

  1. Create tiles out of large images using the Image Tiling Module.
  2. Quickly balance image classes by simple rotate and mirror image augmentation.
  3. Normalize image colors by histogram matching.
  4. Filter histological tiles by white-space presence.
  5. Augment or pre-process images using a plethora of image processing modules.

This release, while mostly stable, will have bugs and unexpected crashes. There will be several updates to the application in the coming weeks. Any suggestions or requests for features are highly encouraged.