StarDist_Fluorescent_cells

This repository includes a StarDist deep learning model and its training and validation datasets for detecting fluorescently labeled cancer cells perfused over an endothelial cell monolayer. The model was trained on 66 images labeled with CellTrace and demonstrated high accuracy, achieving an average F1 Score of 0.877. The dataset and the trained model can be used for biomedical image analysis, particularly in cancer research.

Specifications

Model: StarDist for cancer cell detection

Training Dataset:

Number of Images: 66 paired fluorescent microscopy images and label masks

Microscope: Nikon Eclipse Ti2-E, 10x objective

Data Type: Fluorescent microscopy images with manually segmented masks

File Format: TIFF (.tif)

Brightfield Images: 16-bit

Masks: 8-bit

Image Size: 1024 x 1024 pixels (Pixel size: 1.3205 μm)

Training Parameters:

Epochs: 200

Patch Size: 1024 x 1024 pixels

Batch Size: 2

Performance:

Average F1 Score: 0.877

Average IoU: 0.646

Model Training: Conducted using ZeroCostDL4Mic (https://github.com/HenriquesLab/ZeroCostDL4Mic/wiki)

Reference

Fast label-free live imaging reveals key roles of flow dynamics and CD44-HA interaction in cancer cell arrest on endothelial monolayers

Gautier Follain, Sujan Ghimire, Joanna W. Pylvänäinen, Monika Vaitkevičiūtė, Diana Wurzinger, Camilo Guzmán, James RW Conway, Michal Dibus, Sanna Oikari, Kirsi Rilla, Marko Salmi, Johanna Ivaska, Guillaume Jacquemet

bioRxiv 2024.09.30.615654; doi: https://doi.org/10.1101/2024.09.30.615654