SEOUL, June 30 (Reuters) – Is the young woman in a sundress taking in the view from the bridge? That’s the question South Korean researchers and emergency services are working to answer by using artificial intelligence to detect and prevent suicide attempts.
In this case, this is one of the researchers who demonstrates how difficult it can be for human surveillance teams to tell.
But the AI system they are developing has learned behavioral patterns by analyzing data from cameras, sensors and shipment records from rescue services since April 2020, the Seoul Institute of Technology said on Wednesday.
Based on information from hours of CCTV footage and assessments of details such as the person’s reluctance, the AI can then predict a dangerous situation and immediately alert rescue teams, the researcher said. principal Kim Jun-chul.
“We believe the new video surveillance will allow our teams to detect cases a bit faster and help us respond to a call more quickly,” Kim Hyeong-gil, head of the Yeouido Water Rescue Squad, told Reuters. -temporal images of the Han River bridges in Seoul.
Kim’s team worked with researchers to develop the technology that his team and the Seoul Fire and Disaster Headquarters will pilot from October.
Their work cannot come quickly enough.
South Korea, with a population of 52 million in 2019, had the highest suicide rate in the OECD. More than 13,700 people committed suicide in the same year, according to government data.
Nearly 500 suicide attempts are reported annually on 27 bridges across the nearly 500 km (300 miles) Han River, the city said.
The number of relief shipments increased by around 30% in 2020 from the previous year and many attempts were made by people in their 20s and 30s as the coronavirus pandemic resulted in larger economic hardship and heightened the battle for jobs, Kim of the rescue squad mentioned.
“The system learns the images itself, which can improve results by drastically reducing false alarms,” said the principal investigator.
Minwoo Park reporting, Dogyun Kim; Writing of Sangmi Cha; Editing by Alison Williams
Our Standards: Thomson Reuters Trust Principles.