Researchers are developing a machine learning tool that could one day identify individual animals from photos of their footprints. While some animals, like certain penguin populations, are easy to study due to their lack of fear of humans, others, like Namibia’s brown hyenas, are much more elusive. These nocturnal creatures often avoid even the most cleverly placed camera traps, sometimes destroying them outright.

Marie Lemerle, a researcher with the Brown Hyena Research Project, experienced this firsthand when hyena cubs chewed through her cameras. However, a new collaboration with the US-based nonprofit WildTrack has given her hope. WildTrack, led by British conservationist Zoe Jewell, has been developing an AI-powered system to identify animals from their footprints. Inspired by Jewell’s work tracking black rhinoceroses in Zimbabwe, the tool can currently identify 17 different animals, including leopards, lions, and rhinos. The goal is to refine the system to identify individual animals by their unique footprints.

For the past five months, Lemerle has been collecting hyena tracks at Baker’s Bay, a breeding ground for Cape fur seals on Namibia’s Atlantic coast. She photographs each clear footprint with a ruler for scale, and the WildTrack team at Duke University analyzes the prints in detail. The AI software compares 120 different measurements from each print to find matches, sometimes distinguishing hyenas by subtle differences in toe angles.

Injuries and life experiences also leave unique marks on hyena tracks. Brown hyenas often sustain injuries from fights with other hyenas and jackals, which can affect their footprints. Lemerle believes these differences will help the AI tool identify individual hyenas.

This AI tool could complement traditional study methods, allowing researchers to track hyenas without direct observation. It could also help protect brown hyenas and other threatened species by providing evidence that tracks near farms are not from hyenas, potentially reducing retaliatory killings.

Wesley Gush, a graduate student at the University of Pretoria, believes the tool could significantly aid wildlife researchers and managers. With fewer than 3,000 adult brown hyenas in Namibia and less than 10,000 across southern Africa, this technology could be crucial for their conservation.