Cell cycle phases classification

Download data from the link: https://github.com/ntumitolab/cell-cycle-chirong/releases/tag/publication

System requirement and enviroment settings

A graphics card with at least 6GB of ram.

• Create virtual environment

conda env create -f cellcycleenv.yml

Train an ResNet34 model

• Activate the environment

conda activate cc

• Open Jupyter Notebook (or jupyterlab)

jupyter notebook

• After the Jupyter Notebook has opened, clicking train_ResNet34.ipynb.
• Modify the paths to your data or folder.
• Modify the hyperparameters if you want.
• Start training.

Calssify cell cylce phases by a trained ResNet34 model.

• Activate the environment

conda activate cc

• Open Jupyter Notebook (or jupyterlab)

jupyter notebook

• After the Jupyter Notebook has opened,
  • if the cells weren't labeled, clicking ResNet34_not_labeled_images_classification.
  • if the cells were labeled, clicking ResNet34_labeled_images_classification.
• Modify the paths to your data or folder.
• Start classifying.

Train an SVM model

• Analyze mitochondria
  • Start up ImageJ
  • Open mito_skeleton.ijm
  • Modify the paths
  • Follow the order to Analyze mitochondria
• Open SVM.ipynb.
• Modify the paths to the folders.
• Modify hyperparameters.
• Start training.

Calssify cell cylce phases by a trained SVM model.

• Analyze mitochondria
• Open SVM_classifying_labeled_images.ipynb or SVM_notlabeled.ipynb
• Modify the paths to the folders.
• Start classifying.