This repository is intented to automate the analysis of microscopy fluorescence images of in vitro models of liver-stage malaria. It provides the following functionalities:

1. Segmentation of parasites and hepatocyte nuclei present in the images.
2. Extraction (and plotting) of features informative of parasite- and nucleus morphology.
3. Training a variational autoencoder on the extracted features and assessing distances between experimental groups by calculating the Silhouette coefficient within the latent space.

Examplary Jupyter notebooks are provided under tutorials. Segmentation models were trained on images containing the HSP70- and DAPI stainings at a resolution of 1392x1040.

The recommended way to run this pipeline is by creating a Python 3.9 Anaconda environment and installing the packages specified in requirements.txt.