Deepometry

Deep learning-based image classification and featurization for imaging (flow) cytometry.

This workflow was originally built for imaging flow cytometry data but can be readily adapted for microscopic images of isolated single objects. The modified implementation of ResNet50 allows researchers to use any image frame size and any number of color channels.

Installation

A full installation guide can be found here. Briefly, the following dependencies are needed:

• Python 3.6
• Tensorflow-gpu 1.9.0
• Keras 2.1.5
• Numpy 1.18.1
• Scipy 1.4.1
• Keras-resnet 0.0.7
• Java JDK 8.0 or 11.0
• Python-bioformats 1.5.2

Once the above dependencies are installed, clone this Deepometry repository by :

git clone https://github.com/broadinstitute/deepometry.git
cd deepometry
pip install .

If you want to install deepometry in development mode, run:

pip install --editable .[development]

Use

Execute Deepometry functions through any of the following interfaces:

CLI

Switch to CLI branch:

git checkout CLI

Display a list of available subcommands:

deepometry --help

To display subcommand use and options:

deepometry SUBCOMMAND --help

IPYNB

Switch to IPYNB branch:

git checkout IPYNB

Use these Jupyter notebooks.

GUI (recommended)

Switch to GUI branch:

git checkout GUI

python Deepometry_GUI.py

Open a web-browser, navigate to http://127.0.0.1:5000/ or http://localhost:5000/

Publications

Doan M, Sebastian JA, Caicedo JC, et al. Objective assessment of stored blood quality by deep learning. Proc Natl Acad Sci U S A. 2020;117(35):21381-21390. doi:10.1073/pnas.2001227117

Doan M, Case M, Masic D, et al. Label-Free Leukemia Monitoring by Computer Vision. Cytometry A. 2020;97(4):407-414. doi:10.1002/cyto.a.23987