Researchers Use AI to Track Belugas in Drone Images
Boca Raton, Florida – Researchers at Florida Atlantic University have developed an artificial intelligence method that can detect and track beluga whales in aerial drone images with far less manual labeling than usual.
A video released with the study shows beluga whales in a High Arctic bay. The work, published in Frontiers in Marine Science and announced on November 19, was done with Fisheries and Oceans Canada and the University of Manitoba.
The team trained models on 7,655 aerial images in which whales were marked as adults or calves. Their semi-supervised system, SEMI-DETR, reached useful detection accuracy with only a small share of the images labeled by people, and the improvement was especially clear for calves, which are smaller and easier to miss.
"Instead of requiring researchers to painstakingly label thousands of images before AI can become useful, semi-supervised learning allows the technology to learn from a small amount of expert guidance and then make use of the much larger body of unlabeled imagery," said senior author Hanqi Zhuang, a professor at Florida Atlantic.
Clean HD download
Register a media account or sign in to buy & download this story
Sign in to download Apply for media access