Tiny Gaussian Splatting in 2D

Splatting 2D Gaussians to represent images. Tiny, self-contained, low-overhead, and minimal implementation in a single Jupyter notebook.

This code grew out of some of my research I was doing during my Master's studies. I thought it'd be fun to share a simplified version as a standalone notebook.

I helped create a modified version of this notebook as part of a homework assignment for the Georgia Tech course CS8803/4803 CGA: Computer Graphics in the AI Era. You can find more information about the course here: https://cgai-gatech.vercel.app/. You can also find my corresponding guest lecture on 3D Gaussian Splatting on YouTube: https://youtu.be/MBVmQSA24Yk.

Install dependencies

TLDR: pip install -r requirements.txt

Using pip and venv:

1. Create a virtual environment: python3 -m venv venv
2. Activate this environment:

# MacOS/Linux
source venv/bin/activate

# Windows
venv\Scripts\activate

Once activated, pip will refer to packages installed in the venv folder (e.g. venv/lib/python[your-version]/site-packages)

3. Install required packages: pip install -r requirements.txt

4. (Optional) Include specific version numbers from backup-requirements-frozen-versions.txt

5. You can deactivate the virtual environment using deactivate

Run Jupyter notebook locally

jupyter notebook