This repository contains a pipeline for the analysis of mitochondria morphology using semantic segmentation on Transmission Electron Microscopy (TEM) images. The pipeline includes tools for preprocessing TEM images, training U-Net models for semantic segmentation, generating predictions on new images, and analyzing organelle morphology.
Arriojas Maldonado, A. A., Baek, M., Berner, M. J., Zhurkevich, A., Hinton, Jr., A., Meyer, M., Dobrolecki, L., Lewis, M. T., Zarringhalam, K., & Echeverria, G. (2025). TEM Mitochondria Segmentation Dataset for Triple Negative Breast Cancer Chemotherapy Analysis (1.0.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.15602048
Arriojas Maldonado, A. A., Baek, M., Berner, M. J., Zhurkevich, A., Hinton, Jr., A., Meyer, M., Dobrolecki, L., Lewis, M. T., Zarringhalam, K., & Echeverria, G. (2025). U-Net Model Weights for TEM Mitochondria Segmentation in Triple Negative Breast Cancer (1.0.0). Zenodo. https://doi.org/10.5281/zenodo.15602446
TODO
TODO