VigilantGantry is an open source implementation of an AI-driven automated temperature screening gantry that augments existing thermal systems in use, to enhance the rate of contactless screening. The software is designed to address the limitations of current temperature screening systems and staff fatigue in the temperature screening process.
The intention of the project is to provide automated access control and thermal scanning to limit exposure of frontline staff to large numbers of human traffic flow, while still ensuring that users of the gantry are adequately checked.
VigilantGantry is useful for automating high traffic volume sites screening in detection of symptomatic COVID-19 cases. It aids ground staff to stay vigilant against COVID-19.
We are releasing the face segmentation algorithm that the VigilantGantry uses to identify potential missed out causes by current thermal systems (due to occlusion from fringe, cap, head-gear).
docker build . -t face_seg
docker run --rm --runtime=nvidia -ti -e DISPLAY=$DISPLAY -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix -e QT_X11_NO_MITSHM=1 --name face_seg face_seg:latestHere are suggestions for the open-source community to expand the functionalities:
- Integrate with and test on other thermal systems and thermal cameras
- Integrate with and test on other gantries and door access systems
- Integrate with self-declaration apps for contact tracing purposes
- Integrate with facial recognition technology for contact tracing purposes
- Expand the video analytics features such as detection of coughing, running nose, unwell symptoms, mask on/off, etc.
Find out more about GovTech Data Science & Artificial Intelligence Division at our Medium blog.