shrimpville

Installation

1. Clone sam2 from the repository.

git clone https://github.com/facebookresearch/sam2.git && cd sam2

2. Add path to sam2 to environment (i.e add to bashrc or zshrc)

echo "export SAM2_PATH=\"//home/$USER/sam2/\"" >> ~/.zshrc && source ~/.zshrc

or for bash

echo "export SAM2_PATH=\"//home/$USER/sam2/\"" >> ~/.bashrc && source ~/.bashrc

3. Create virtual environment for python

python -m venv py-venv

4. Activate virt environment

source py-venv/bin/activate

5. Installing sam2 dependencies with the virtual python environment

pip3 install -e . && \
pip3 install -e ".[notebooks]"

6. Install model checkpoints

cd checkpoints && \
./download_ckpts.sh && \
cd ..

7. Finish installing dependencies used outside of sam2 requirements

pip3 install pandas

To deactivate the python virtual environment, simply run:

deactivate

Usage

1. To run any of the python scripts, activate the python virtual environment:

source py-venv/bin/activate

2. Then if you need your video split into individaul frames run:

./extractFramesFromVid.sh videos/<file of video>

3. After this to get coordinates to start with run:

./extractCoord.py videos/<dir of jpgs created from previous script>

4. Then finaly run the main script to get csv

./extractMasks.py videos/<dir of jpgs created from previous script>

If you want to get the csv points ploted onto a grid to see the range of movement for the object simply run

./trackmotion.py videos/<dir of jpgs created from previous script>

To deactivate the python virtual environment, simply run:

deactivate