COVIDPreventionSystem

Computer vision applied to COVID prevention techniques, developed for the thesis work by Davide Gena of the University of Calabria 2020/2021.

Main library used

The OpenCV library allows to recognize people in the video source by model of different neural deep network frameworks.

In this case it was used the Mobilenet-SSD model of the Caffe framework.

OpenVINO Toolkit

The performance can be increased with OpenVINO, a free toolkit developed by Intel to speed up the work of the neural network. To use it on this project you need to download and install the OpenVINO engine on your machine and uncomment two line of code in the "PreventionDetectors" class.

Go Here to download the toolkit.

To run it open in the same terminal the toolkit and after run the executable or your IDE, like PyCharm. To run the toolkit:

source /opt/intel/openvino_2021/bin/setupvars.sh

to run your IDE (PyCharm in this case):

cd pycharm-XXXX.X/bin/
./pycharm.sh

Interface

The interface, that allow to configure the system, was developed with the PySimpleGUI library. This library simplifies many others graphics framework and work by default with Tkinter.

Here are some examples of the interface and system operation.

Main interface:

Examples of gathering detection:

Example of mvement detection:

A portion of log file: