Japan is planning to use facial recognition to identify so-called “troublemakers” among its community. With increasing bear sightings in urban areas across the country, and a rise in the number of attacks, the town of Shibetsu is hoping that AI would enable it to manage the situation better and keep people safe, reports Mainichi Shimbun. 

While all bears look the same from a distance, small differences in appearance like the distance between the eyes and nose let facial recognition work properly on them.

Images required to train system

To get the system working, the tech needs at least 30 images of each bear’s face taken from the front. Automatic cameras have been installed by the workers at the South Shiretoko Brown Bear Information Center. The cameras are placed at known bear trails to gather essential data, but so far it hasn’t managed to gather enough images to run facial recognition.

The chances of getting a bear to look straight into the camera are very slim, but the team is hoping that it will soon have enough images to launch the program. The aim is to help workers understand the specific behavior traits of each bear and capture the ones that could be a potential threat to nearby villagers.

Shortage of food in natural habitat

This is not the first time such a technology has been used on bears. Researchers in the U.S. and Canada have tried out similar systems with an aim to gauge the population in forest reserves. 

Bear attacks in Japan have been making headlines. Recently, a bear injured four people in Hokkaido’s capital city of Sapporo before it was shot dead. The footage of the attack showed the bear striking from behind.

Back in 2019, Japan witnessed around 150 bear attacks, which was the highest number of incidents in a decade. Experts believe the increase in such cases could be due to a shortage of food in the bears’ natural habitat.