This repository stores all materials for a workshop on How to build your motion tracking pipeline which takes place on the 12th and 13th of March at ZAS (Berlin).
To make sure we all start on the same page, please prepare your laptop by following the instructions below.
To be able to follow the workshop and explore the coding procedures of building a motion tracking pipeline, you will need to have a Python available.
The easiest way to get Python is installing Anaconda, a common platform for this programming language.
Important
At some point during the installation process, you need to check ☑️ that you want to add Anaconda/Python to system path. Without that, your command prompt or the text editor will not know where to find executor.
If you feel you need a guidance, you can follow one of many tutorials online (e.g., here for Mac, and here for Windows).
To work with Python scripts, it is very convenient to have a text editor. Essentially, this can be any software that can edit texts, but I personally recommend - and for this workshop highly encourage - to use Visual Studio Code.
Once you have installed it, you need to do few more adjustments so that VS Code can recognize a Python script. Go to Extension panel (Ctrl + Shift + X) and search for Python. Click on the install button, et voilà! Do the same with extension Jupyter.
Tip
One of the reasons why VS Code is really nice lies in the extensions. One of my favorites is Data Wrangle that allows you to look through your dataframe in much more friendly way than, for instance, RStudio (no hate though 👼)
Once you are all set up, it's time to get this repository to your PC.
Here, you can either use Github Desktop which will later also allow you to synchronize your local version of this repository with any updates I may do here.
Note
How to clone repository via Github? After installing, launching Github Desktop, and signing in with your Github credentials, simply click on File > Clone repository... > URL and paste there the link to this repository, https://github.com/sarkadava/MotionTrackingPipeline_Workshop_ZAS2025
If you don't want to install yet another software, you can just download the repository. At the top of this page, click on <> Code and then Download ZIP. This will download the repository.
During the workshop, we will work with so-called virtual environment. I will not go into details here, but environment essentially make sure that our project is safely isolated from the rest of the Python packages and certain versions therefore do not collide.
Now turn on your Anaconda Prompt and type
cd C:/Directory/where_the_downloaded/folder_lives
Caution
If you are a Mac user, you might notice that there is nothing like Anaconda Prompt to be found. You can simply use your Terminal.
Now it's time for our environment. You can create one by typing
conda create --name MTworkshop python=3.12
and then activate it with
conda activate MTworkshop
You should now see (MTworkshop) in the beginning of your line.
Finally, we want to install some packages that I have prepared into the requirements.txt file. In your prompt, type
pip install -r requirements.txt
Now you can open your Visual Studio Code, click on File > Open Folder and select the folder where this repository lives. In the right panel, you should be able to see all the files. Click on the jupyter notebook Test.ipynb
In the right-down corner, you might see a prompt {select an Interpreter}. If you do, click on it and select 'base'.
Click on the play button. A drop-down menu will pop up to select a kernel source. Click on Select Python Environment and now choose the one we have created a second ago, MTworkshop. If it worked, you should be able to see the name of the kernel in the top-right corner like so
Caution
If you are don't see Select Python Environment but instead VS Code suggests an option to recommend and install needed extensions - click on it. You probably did not install Jupyter extension in the previous steps and VS Code will now do it for you (another reason why to like this smart editor 👼).
When you click on the play button, most likely a window will pop up, asking you to install ipykernel on your VS Code. Click on install and wait until it finishes.
Now, finally, the moment of truth has come! If nothing screams red and a welcome message got printed, congratulations, you are ready to go! 😊
Unfortunately, there are many things on the way that can go wrong - for many reasons. Yes, technology can be too stubborn. While it is impossible to cover all of them, here are the most common errors you may encounter.
This error occurs when your system does not know where to find the pip command. Pip is either not installed in your environment, or it is not added to the system path.
You can install pip by typing
conda install pip
If it does not work, refer to the next solution.
Go to system settings and search for Environment Variables. Click on Environment Variables and in the System variables find Path. Click on Edit and then New. Now you need to add the path to your Anaconda. This is usually in C:\Users\YourName\anaconda3\python.exe.
If you don't see the environment you have created in the list of available kernels, but are 100% sure that you have created it, go to your Anaconda Prompt and type
python -m ipykernel install --user --name=MTworkshop
Now you should see it in the list of available kernels. Note that maybe you will need to restart your VS Code.
If during the Step 5 you get this error (or some other module is missing), it means that the package was not installed.
You can try either reinstalling the whole requirements
pip install -r requirements.txt
Important
Make sure that you are in your MTworkshop environment. It should be displayed in brackets at the beginning of your prompt.
Or just install the missing package
pip install mediapipe
If you have encountered a problem that is not listed here, please let me know - ideally before Wednesday. I will do my best to help you out. You can reach me via email kadava[et]leibniz-zas[dot]de