Is the young woman in summer just enjoying the view on the bridge? This is a question that South Korean researchers and emergency services are trying to use artificial intelligence to detect and prevent suicide attempts to answer.

In this case, one of the researchers demonstrated the difficulty of the human surveillance team’s judgment.

But Seoul Institute of Technology said on Wednesday that since April 2020, the artificial intelligence system they are developing has been learning behavior patterns by analyzing data from cameras, sensors, and dispatch records of rescue services.

Lead researcher Kim Jun-chul said that based on hours of closed-circuit television video information and the hesitation of assessors and other details, artificial intelligence can predict dangerous situations and immediately alert rescue teams.

“We believe that the new closed-circuit television will enable our staff to detect cases faster and help us receive calls more quickly,” Kim Hyung-gil, who is in charge of the Yeouido Water Rescue Team, told Reuters that he monitored the real situation. . -Time footage of the bridge on the Han River in Seoul.

Kim’s team has been working with researchers to come up with a technology that his team and the Seoul Fire and Disaster Headquarters will try out from October.

Their work didn’t come fast enough.

With a population of 52 million in 2019, South Korea has the highest suicide rate in the OECD. Government statistics show that more than 13,700 people committed suicide in the same year.

The city stated that nearly 500 suicide attempts are reported every year on 27 bridges over the nearly 500-kilometer (300-mile) Han River.

See also  Apple says child abuse detection system will check photos in iCloud

Compared with the previous year, the number of rescue dispatches in 2020 has surged by about 30%. Many attempts have been made by people in their 20s and 30s because the coronavirus pandemic has brought greater economic difficulties. And it has intensified the competition for employment, said Kim of the rescue team.

“The system learns the lens by itself, which can improve results by greatly reducing false alarms,” ​​said the lead researcher.

© Thomson Reuters 2021