Manage episode 282599836 series 2530089
著作 レアジョブ英会話 の情報はPlayer FM及びコミュニティによって発見されました。著作権は出版社によって所持されます。そして、番組のオーディオは、その出版社のサーバから直接にストリーミングされます。Player FMで購読ボタンをタップし、更新できて、または他のポッドキャストアプリにフィードのURLを貼り付けます。
A new facial recognition program can now identify bears individually from each other. Tracking bears is important to scientists because it helps in research and conservation of the animals. However, this is no easy task since bears do not have unique identifiers that people can use to tell them apart. The animals also get bigger and smaller depending on the season, and their looks frequently change during their lifespans. To address this issue, bear biologist Melanie Clapham teamed up with two software developers to create BearID, a project that uses facial recognition software to monitor grizzly bears. Together, the team snapped 4,674 photos of grizzly bears from bear-viewing areas in Alaska and British Columbia. Then, they used an artificial intelligence (AI) software that can recognize bear faces in the images. It measured the proximity between the animals’ eyes, nose, ears, and forehead to recognize them. The software was able to correctly identify 132 of the animals individually. With an 84% accuracy rate, Clapham said that BearID is a cheaper, less invasive, and safer way to track animals compared to electronic tags and collars. She hopes that different organizations will use BearID to observe animal behavior and the movements of endangered animals. Despite the project’s positive results, wildlife expert Tanya Berger-Wolf believes that the technology could be used against animals. She said that access to the animals’ data must be controlled because if poachers manage to get hold of the bears’ images, they could track the animals’ locations and hunt them using BearID.