Engineer Ahmed Manaa Al-Attar, a PhD student in robotics and artificial intelligence engineering at Imperial College London, revealed the success of a group of engineers in employing artificial intelligence technology in monitoring and detecting compliance with novel Coronavirus (COVID-19) preventive measures by monitoring the wearers of masks using a deep learning system.
"Twelve youth (engineers and students of engineering majors) have innovated an automated system to detect mask wearing violations, through the use of artificial intelligence technology, to facilitate the application of preventive measures in malls, airports, ports, schools, universities, and workplaces". Al-Attar expressed. "Innovation is a result of cooperation between Society of Engineers' Youth Engineers Council and the Association of Emirati Engineers Abroad, with the participation of a group of distinguished engineers and students from different universities in the UAE in innovation and implementation of the 100% remote engineering project, aimed at employing artificial intelligence technologies to solve challenges and problems that may arise with lifting restrictions on movement and the return of economic activity." Al-Attar noted.
The innovated tool consists of a thermal camera that captures an image of the face, analyzes the image, and releases an alarm if the mask is not worn. This process does not exceed three seconds by using an intelligent algorithm.
"The developed technology processes the image of the person, then enters the image into a smart algorithm that classifies it into whether mask wearing or non-wearing. In the event that the person is a non-wearing of mask, the technology releases an alarm or inform the regulatory authorities, which enhances the preventive measures to combat «COVID-19 » in Public places and malls." Al-Attar clarified.
For her part, Eng. Maryam Al Marzouqi, member of the Society of Engineers' Youth Engineers Council, stated that artificial intelligence techniques are able to deal and process large-size images, which reduces human intervention and saves time and effort to analyze images, as the smart algorithm works to detect the mask wearing or non-wearing. "We have overcome the challenges we faced in analyzing images by inserting specific image sizes, external effects such as lighting and background, as well as the color of the mask to reach to the highest accuracy for programming and results.
While the team member, Eng. Muhammad Osama Al-Noor, revealed the team's use of the thermal map to visualize the mechanism of smart algorithm capturing of the mask wearer. The thermal map represents the areas around which the algorithm revolves while monitoring the mask, and through analyzing the images mask non-wearer is correctly detected.
Neural Network
"Building the neural network was a new challenge, as the team took charge of the project programming in a very short time, while maintaining a high accuracy percentage." Member of the innovation team, Eng. Mohammed Saeed Al-Hajjar, stated.
"The project helped us discover new horizons for scientific developments that can be used to overcome the current crisis." He pointed out.