Skip to content

Package using StarDist and Python that performs object detection and spatial analysis on H&E images

License

Notifications You must be signed in to change notification settings

ajinkya-kulkarni/PySpatialHistologyAnalysis

Repository files navigation

Streamlit App License: GPL v3 DOI GitHub commit activity GitHub release (latest by date)

Demonstrating PySpatialHistologyAnalysis

App Overview

This is primarily a web application for analyzing H&E stained images using the PySpatialHistologyAnalysis package, which utilizes the StarDist packages under it's hood. The application allows the user to upload an H&E image. It first stain normalizes the image and then performs object (nuclei) detection on the image using the StarDist2D model. The detected objects are then highlighted and displayed alongside the original image. Spatial analysis is then performed on the detected nuclei, and finally a spreadsheet of all the results is displayed.

Dependencies

The app is built using the Streamlit framework and requires the dependencies as mentioned in the requirements.txt file.

Using the app

App Screenshots App Screenshots App Screenshots App Screenshots

To use the app, simply upload an H&E image using the file upload widget and click the "Analyze" button. The app will then perform object detection on the image using the StarDist2D model and display the results alongside the original image. If an error occurs during image analysis, an error message will be displayed. Note that the app works best for images smaller than 1000x1000 pixels.

References:

  1. All images taken from Link 1 and Link 2