BacTrak is a python implementation of the tracking algorithm described in the following manuscript:
S. Sarmadi et al. Stochastic Neural Networks for Automatic Cell Tracking in Microscopy Image Sequences of Bacterial Colonies. Mathematical and Computational Applications, 27(2):22, 2022 [arxiv, doi].
If you use this code we would appreciate if you add the above reference to your citations.
We have included some benchmark files into this repository: Two binary images named 'J.png' (first frame) and 'J_plus.png' (second frame) in the Sample directory. These images are shown below. The software used to generate these synthetic images can be found in the following repository: https://github.com/jwinkle/eQ. We have included two synthetic data sets for testing our code:
To execute the code simply go to the toplevel directory and run the python script:
python3 BacTrak.py
After executing the code the following information will be displayed in your command window:
Linking mother and children started
11%|█ | 551/5000 [00:26<03:35, 20.64it/s]
No split registration started
20%|█▉ | 394/2000 [27:41<1:52:53, 4.22s/it]
At the end the code will return a python dictionary which is the result of the registration of cells. The dictionary will be saved as Registration.npy in the folder Sample. The keys of the dictionary are the labels of the cells in 'J' and the values of the dictionary are the labels of the associated cells in 'J_plus'. If a cell devides we have a tuple for the value in the dictionary.
Besides that, we also store pictures to illustrate the registration results inside the Sample folder. The colors of the cells designate the identified mapping. The colors establish a cell-to-cell correspondence between the cells in the image 'J' and the image 'J_plus'. We include an illustration below.
Our software depends on the following python packages:
OpenCV(installation:pip install opencv-contrib-python; https://pypi.org/project/opencv-python)scikit-image(installation:pip install scikit-image; https://scikit-image.org)statsmodels(installation:pip install statsmodels; https://www.statsmodels.org/stable/index.html)tqdm(installation:pip install tqdm; https://pypi.org/project/tqdm)
A BibTeX entry for LaTeX users is
@article{Sarmadi:2022a,
author = "S. Sarmadi and J. J. Winkle and R. N. Alnahhas and M. R. Bennett and K. Josic and A. Mang and R. Azencott"
title = "Stochastic Neural Networks for Automatic Cell Tracking in Microscopy Image Sequences of Bacterial Colonies",
journal = "Mathematical and Computational Applications",
volume = "27",
number = "2",
pages = "22",
year = "2022"}