viraatdas / ColonyCounter

count colonies from an image

Colony Counter

Count colonies from petri dish images.

Sample outputs

Colony Counter Results

The following table shows the input colony images and their corresponding output images with the colony count and superimposed points:

Input Image	Output Image

Usage

To use the Colony Counter, first, make sure you have Python installed on your system. Then, follow these steps:

cd backend/

1. Install the required dependencies using pip:

   pip install -r requirements.txt

2. Clone the repository

   git clone https://github.com/viraatdas/ColonyCounter.git

3. Navigate to the project directory

   cd ColonyCounter/

4. Run the colony counter script with the path to the input image as an argument:

   python count.py sample-images/input/sample_1.png

This will process the input image, count the colonies, and generate an output image displaying the detected colonies along with the colony count in a window.

Image parsing logic

fill this more

Canny edge detection

Backend

• Deployed as a docker image to Heroku
• Flask API

Frontend

• React
• Deployed on netlify

Todo (convert these to GitHub issues)

• Improve the algorithm to achieve higher accuracy
• Use manual validation to check if it works ona diverse dataset
• Add batch capability to process multiple files
• Convert to webapp to make this more accessible
  • Maybe create a lambda that is accessible as an API

Contributing

PRs are always welcome for any improvement

Citation

If you use this Colony Counter in your research work or project and want to cite, you can use the following BibTex entry:

@misc{colony-counter,
title={Colony Counter},
author={Viraat Das},
year={2024},
howpublished={\url{https://github.com/viraatdas/ColonyCounter}},
}