Ultrack Workshop I2K 2023

Ultrack from Images to Knowledge - 2023 workshop

There are two options to follow this tutorial, locally or with Google Colab.

Local setup

You must have conda or mamba installed. If you don't have it, you can follow their instructions.

First, clone this repository:

git clone https://github.com/royerlab/ultrack-i2k2023

Go to the repository folder:

cd ultrack-i2k2023

Then, you can create a new environment with the following command:

conda env create --file environment.yml

NOTE: If you're using apple silicon, you must use the environment-apple.yml file instead.

This will create a new environment called ultrack-i2k2023. You can activate it with:

conda activate ultrack-i2k2023

Then, you can browse the notebooks with:

jupyter lab

Google Colab

For Google Colab, you can follow the links below. You will need a Google account.

To access GPUs on Google Colab, you can go to Runtime > Change runtime type and select GPU as the hardware accelerator.

• Introduction:

  https://colab.research.google.com/github/royerlab/ultrack-i2k2023/blob/main/intro.ipynb

• Multiple hypotheses tracking:

  https://colab.research.google.com/github/royerlab/ultrack-i2k2023/blob/main/multiple-labels.ipynb

Additional Resources

For best performance we recommend using gurobi. It's a commercial software, but it's free for academic use. You can get a license here.