Project Team: Erich Choudhury, Connor Bowler, Alex Wirtz Supervisor: Dr. Hamed Tabkhi Date: May 13, 2021
This is the final project for Real Time AI at UNCC, Spring 2021. This implementation is based off of Mikel Brostrom's YOLOv5_DeepSort_Pytorch project.
This project is a near real time COVID-19 social distance monitoring system the uses pre-existing libraries to help with estimating distances between objects and verifying social distancing guideline in order to create a safer, healthier public. It will also assist in visualizing if two people are too close together.
To run, first run install the required libraries from requirements.txt with:
pip install -r requirements.txt
You will need to install python dependencies for YOLOv5, Deep Sort, TensorFlow, and Pytorch. This is described in the README for each project. Additionally, you will need to download the NYU Depth Dataset as described in the README for FCRN. Place it in the root directory and then run the following command from a static location.:
python predictdepth.py NYU_FCRN.ckpt img.jpg
Finally, run continuously with:
python covidtrackerdepth.py --source 0