2@DBE'2021 AI TECH Analysis of infrared images and segmentation of facial features on thermograms using SI for the purposes of COVID-19 prevention

project type: application

edition: 2021

number of students in project 3 - 5

manager: Piotr Filipowicz

The aim of the project is to implement the system application module for screening temperature monitoring of people entering the room. The module is to be used to automatically designate the area of ​​the face and characteristic elements of the face on thermograms, taking into account the distance of the person from the thermographic camera. It is assumed that people enter the room from the outside, where the weather conditions change, which results in different temperature distributions on the facial features. In addition, the person appears in the frame even in the far perspective and the software, depending on the distance of the person from the camera, would determine the elements of the face taken to determine the body temperature. Segmentation methods, optimized for thermograms, and machine learning methods, known from the literature, would be used for segmentation.

 

Tasks:

Analysis of the collected thermographic sequences throughout 2020 - preparation of a report with conclusions on the average face temperature depending on the season and weather conditions.

Analysis of the state of knowledge and existing solutions for monitoring people for fever.

Analysis of SI methods in image segmentation, including thermograms.

Defining design assumptions and algorithms for tracking the position of the face on thermograms and determining the characteristics of the face.

Application module implementation.

Tests and verification on thermographic sequences recorded in various weather conditions.

Preparation of the final report