OAK Module
OAK module that groups all the scripts and functions necessary for neonatal infant monitoring.
Environment configuration
Execute with conda installed:
conda env create -f env.yml
conda activate oak
pip install -r requirements.txt
If new libraries are installed, it is necessary to execute the following command:
pip-compile --extra=dev
Structure
Scripts
With the OAK camera plugged into your device, and the oak environment active (see Environment configuration) run the following command from the root folder:
python demo_video.py
The script will load the session stored in the demo folder. This session last 20 seconds and it consist on three videos, each recorded by a different camera of the OAK device. A window will appear on screen, showing the video from the center camera. When the neonate is detected, the script draws a bounding box around their face, and other surrounding the body. The script also plots three graphs next to the image. Each graph shows a different score on real-time:
- Stress score (between 0 and 1)
- Limb stretch scores (between 0 and 1)
- Breath
Instead of feeding a video to the script, you can record on real-time:
python demo_cam.py
To stop the record, press "q" then "Ctrl + C".
Raspberry installation
First install all libraries:
python3.7 -m pip install -r requirements.txt
python3.7 -m pip install -e .
Now install important cv2 dependencies:
sudo apt upgrade
sudo apt-get install -y libatlas-base-dev libhdf5-dev libhdf5-serial-dev libatlas-base-dev libjasper-dev libqtgui4 libqt4-test