Development of face mask detection system using machine vision

Chan, Yoke Lin (2023) Development of face mask detection system using machine vision. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Development of face mask detection system using machine vision.pdf - Submitted Version Download (486kB)
	Text (Full text) Development of face mask detection system using machine vision.pdf - Submitted Version Restricted to Registered users only Download (2MB)

Abstract

Coronavirus disease (COVID-19) spreads in tiny liquid particles from mouth or nose. To prevent people from being infected by Covid-19, the Malaysia government has made it compulsory for people to wear masks in public since 1st Aug 2020. This is difficult to make sure everyone follows the rule especially when people came in a crowd. Hence, face mask detection system is proposed to identify whether a person is wearing mask or not. The scope of this project is limited to white surgical mask and front view of the face. This system is created using MATLAB with computer vision toolbox. The method used is training Cascade Object detector. This is a feature based detector and need to be trained. Cross-validation is used to train the detector due to small dataset. After testing, accuracy of the system is calculated from the output included positive and negative images. In this project, the accuracy for first training with the original size images is 76.67%, second training with the cropped ratio 9:11 is 67.50% and accuracy for third training process with 9:4:7 cropped images is 89.17%. In future work, system can be improved to detect other colours and types of face masks.

Actions (login required)

View Item