pyMarAI

Collection of neural network models for automatic image segmentation of microscopic tumor spheroids. Intended to be used with nnU-Net deep-learning framework. Trained and tested on a total of microscopic images of mouse pheochromocytoma (MPC) tumor cells.

In addition to the trained network model, a PyQt5-based graphical user interface tool is provided. This tool provides a complete pipeline for handling microscopic spheroid image data, running deep-learning–based delineation, and curating results for continuous model improvement.

pyMarAI deep-learning tumor spheroid segmentation overview © Jens Maus

Find pyMarAI on Helmholtz Imaging Connect
Authors: Jens Maus, Janina Nitschke, Pavel Nikulin, Frank Hofheinz, Mareike Barth, Sandy Lemm, Lena Richter, Jens Pietzsch, Anja Braune, Martin Ullrich