A major Japanese cosmetics company has started testing an artificial intelligence-based technology that allows users to instantly discover their risk of heatstroke by just presenting their faces to a camera.
An AI-based risk assessment system has been put to the test at construction sites around the country by Pola Orbis Holdings Inc., a company based in Chuo Ward, Tokyo, since June. This is an effort to use the business’s expertise in face and skin research to address a social issue.
The business is working with the Toyota College in Toyota, Aichi Prefecture, the National Institute of Technology, and the IT startup Dumsco Inc. in Tokyo’s Minato Ward. The method for recognizing the risk of heatstroke makes use of AI technologies created by Toyota College.
The AI-powered technology employs facial expression recognition to identify people who may be at risk for heat stroke, such as those who have a temperature of 37.5°C or higher or who have spent three hours or more outside. The temperature and humidity at the moment are taken into consideration to calculate the likelihood of suffering from heatstroke. The creation of cameras and other tools for use on building sites fell under the purview of Dumsco.
In the trial, workers are required to reveal their faces before morning meetings and lunch breaks using tablet computers that have cameras mounted at construction sites in Tokyo and other locations, according to Pola Orbis. A “high risk” or other outcome is presented in roughly three seconds, and the results are also delivered to a terminal device that the site supervisor is carrying. The technology is designed to enable the site manager to direct high-risk employees to take a break.
The Ministry of Health, Labor, and Welfare reports that in 2022, 827 workers suffered from heat stroke at work and either died or missed at least four days of work. Construction sites had the most instances of this type of incident, at 179. Although there are tools to assess the risk of heatstroke, there are problems, such as the challenge of simultaneously identifying several components.