Covid-19, wearing N-95 masks in clinical environments: Thermographic detection of air leaks

Type de document

Articles dans des actes de congrès

Année

2022

Langue

Anglais

Titre des actes

Proceedings of the 2022 International Conference on Quantitative InfraRed Thermography (QIRT 2022)

Maison d’édition

QIRT Council

Résumé

The presence of leaks in N95 masks represents a major issue and increases the risk of contamination. Usually, detecting these leaks is achieved manually and using expensive devices. The methods presented in this paper aim to automatically detect and locate leaks in N95 masks using recorded thermal image sequences. It is achieved by detecting the temperature variations on the skin around the N95 mask. These variations are closely related to leaks and are caused by the breathing. The first proposed approach consists in processing the infrared digital images where traditional methods such as the superposition and subtraction are employed. The second one employs a deep neural network architecture trained with labelled data before performing leak detection.

Mots-clés

COVID-19, SARS-CoV-2, Masque N95, N95 mask, Détection des fuites, Leak detection

Numéro de projet IRSST

2022-0008

Link to Full Text

Partager

COinS