Getting started

Pathadin- bridging pathologist and AI

Pathadin was designed to contribute to spreading quantitative and objective pathology.  It is easy to use, yet powerful combinatory open-source tool, aiming to introduce the main possibilities of modern deep learning solutions and fill the gap between pathologists and developers.

Latest release and system requirements

The latest release and the source code is available:

Here and at GitLab

Minimum system requirements for adequate experience include monitor resolution 1280 × 720, operation system Windows 7 SP1 or newer, CPU Intel DualCore, RAM 4 GB.

Getting started and main tools

As Pathadin is installed, You will see the list of main tools.

Pathadin is constantly evolving. Click the name of the tool to see a detailed description. By now, the main tools include.

  • Add subwindow- for viewing multiple images.
    • Example: viewing the series of sections, parallel viewing H&E and immunohistochemistry.
    • Images can be synchronized (link to sync) in various ways.
  • Tile subwindows.
  • Synchronize- synchronizes slides and annotations.
    • Example: for analyzing the corresponding regions in a series of sections, open slides in multiple windows, press synchronize all and draw an annotation on any of slides- the annotation will be shown on all the synchronized slides.
  • Open image- Pathadin supports OpenSlide library and main image formats.
  • Show properties- display properties of an opened image and metadata of a digital slide.
  • Grid- displays grid. The grid size can be set in pixels or microns.
  • Save screenshot- makes a screenshot in the desired format.
  • Copy to clipboard- for fast copy/pasting an image from Pathadin.
  • Pan/select- moving image and selectin annotations.
  • Annotations– defying ROI for analysis and dataset generation.
    • Line
    • Rectangle
    • Circle
    • Polygon
  • Fit- fits the image to windows size.
  • Magnification size- for digital slides, displays the level of microscopic magnification.

Analysis

The analysis is done inside the annotations.

First, select and define the parameters of the filter to apply (K- means, Nuclear counting, trained AI model, etc) – filter ID. The filter selection is available in the upper right section.

Afterward, the annotation for the filter application must be selected. The list annotation is available in the lower right window. Apply the filter by choosing the filter ID.

The nuclear counting module is provided by HisytmoicsTK; the source code and guidelines are at their homepage.