This repository is no longer maintained. There has been tremendous progress since the release of this code in the field of computer vision and we encourage you to seek more up-to-date alternative implementations. While we haven't tested this, we suggest you look at OrganoID recently released by J. Matthews et al. from the University of Chicago.
OrgaQuant is a simple-to-use python script with a user interface for measuring organoid number and diameter in 3D cultures. While this repository contains an improved version of the code described in our original paper, please refer to the manuscript for details. The original code described in the paper is available upon request but is much slower than this current version. We highly recommend using the updated code available here.
We recommend a computer with at least 16 GB of RAM and an NVIDIA graphics card (required) with at least 8 GB of GPU memory. All code was developed and tested in a Windows environment, but should work fine on both Mac OS and Linux.
- Install Anaconda from https://www.anaconda.com/distribution/
- Open the Anaconda Prompt and create a new conda environment using
conda create -n orgaquant python=3.6 - Activate the newly created environment using
activate orgaquant - Install Tensorflow and Git using
conda install tensorflow-gpu=1.14 git - Install dependencies using
pip install keras-resnet==0.2.0 cython keras matplotlib opencv-python progressbar2 streamlit - Clone the OrgaQuant repository using
git clone https://github.com/TKassis/OrgaQuant.git - Move into the directory using
cd OrgaQuant - Install keras_retinanet using
python setup.py build_ext --inplace. More information here - Download the pre-trained model from here and place inside the trained_models folder.
- Within the OrgaQuant directory run the following:
streamlit run orgaquant.py. This should automatically open a browser window with the user interface. - Indicate the folder that contains the organoid images (default is /test_folder)
- Modify the sliders to adjust some of the settings based on your images.
- You can batch process all your images by clicking on 'Process All'. This will use the same settings you adjusted for the sample image.
When the script finishes running you should have a CSV file with the results as well as a labeled image file for each image in the folder. If you encounter any problems please raise the issue here on Github.
If you have certain types of organoids for which the trained model provided here does not work we might be able to create a new trained model for you and make it available here if you share some of your images with us. Please raise the request as a Github Issue. If you would like to train your own model, you may follow the instructions given for the Keras RetinaNet package.
If you have any improvements in mind please feel free to contribute to the project via GitHub. If you encounter any issue running the above code please raise the issue here.
If you use this code, please cite:
Kassis et al., OrgaQuant: Human Intestinal Organoid Localization and Quantification Using Deep Convolutional Neural Networks. Sci. Rep. 9, 1–7 (2019).
The open access paper can be accessed through https://www.nature.com/articles/s41598-019-48874-y.
OrgaQuant builds on the object detection framework in Focal Loss for Dense Object Detection by Lin et al., 2017.