Implements a basic multi-scale cursor detection using openCV.
- Works across multiple scales.
- Works (at all!) (see below)
To view the processed video, please check this link.
- May not function effectively on non-white backgrounds, though conversion to grayscale could mitigate this issue (results may vary).
- Performance is fucked
Install using uv
and run using python
.
#If you haven't installed UV yet, run the following command:
curl -LsSf https://astral.sh/uv/install.sh | sh
# create a venv with uv
uv venv
# activate venv
source .venv/bin/activate
# install reqs (just opencv)
uv pip install -r requirements.txt
# run the thing
python main.py
# hit space or any key when you are done viewing the image
python video.py
# Or, if you want to parallelize the processing,
# this has a dependency on FFMPEG - if you don't have it installed on your machine, install it
python paralell_video.py
My m2 does a 7 second 2.5k video @ 30 fps in about 8 minutes using the parallelized script