Scientists say facial recognition could help save seals

FREEPORT, Maine (AP) — Facial recognition technology is mostly associated with uses like tracking and authenticating human faces, but scientists believe they’ve found a new use for it — preserving seals.

A research team at Colgate University developed SealNet, a database of seal faces created by photographing dozens of fur seals in Maine’s Casco Bay. The team found that the tool’s accuracy in identifying marine mammals approached 100%, which is no small feat in an ecosystem home to thousands of seals.

The researchers are working to expand their database to make it available to other scientists, said Christa Ingram, a Colgate biology professor and team member. Expanding the database on rare species such as the Mediterranean monk seal and Hawaiian monk seal could help inform conservation efforts for those species, she said.

Cataloging seal faces and using machine learning to identify them could also help scientists get a better idea of ​​where seals are in the ocean, Ingram said.

“Understanding their dispersal, understanding their patterns really helps inform any coastal conservation efforts,” she said. “For mobile marine mammals that move around a lot and are difficult to photograph in the water, we need to be able to identify individuals.”

SealNet is designed to automatically detect a face in an image, crop and recognize it based on facial images such as eyes and nose shape, as it would be in a human. A similar tool called PrimNet, designed for use on primates, had previously been used on seals, but SealNet outperformed it, the Colgate researchers said.

The Colgate team published their findings in April in the scientific journal Ecology and Evolution. They processed more than 1,700 images of more than 400 individual seals, the report said.

The paper states that “the ease and wealth of image data that can be processed with SealNet software is a vital tool for marine mammal ecological and behavioral research in the emerging field of conservation technology.”

Seals are a a US conservation success story The animals were once prized in New England, where fishermen considered them pests in the 19th and early 20th centuries. But the Marine Mammal Protection Act, which turned 50 in October, extended new protections to them — and the population began to recover.

Seals and other marine mammals have long been studied using satellite trackers. Using artificial intelligence to study them is a way to bring conservation into the 21st century, said Jason Holmberg, chief executive of Wild Me, an Oregon-based company working to bring machine learning to biologists. Wild Me is developing a potential partnership with SealNet.

“It’s a shift and an upgrade of ‘big brother’ style technology to a very benign conservation goal,” Holmberg said.

Harbor seals are now quite numerous in New England waters, where they climb onto rocks and delight seal-watching cruises and beachgoers. However, other species of seals remain at risk. The Mediterranean monk seal is considered the most endangered seal in the world, with only a few hundred animals left.

Using facial recognition could provide more valuable data, said Michelle Berger, a research associate at the Shaw Institute in Maine who was not involved in the SealNet study.

“Once the system is perfected, I can imagine many interesting environmental applications for it,” Berger said. “If they could recognize the seals and recognize them from year to year, that would give us a lot of information about movement, how much they move from place to place.”

Colgate researchers are also working with FruitPunch, a Dutch artificial intelligence company, to improve some aspects of SealNet to encourage wider use. FruitPunch is bringing in several dozen scientists from around the world to work on the task of optimizing SealNet’s workflow, said Tjemme Duper, FruitPunch’s head of partnerships and growth.

Improved automation of facial recognition technology could make SealNet more useful to more scientists, Duper said. This will open up new opportunities to study the animals and help protect them, he said.

“It helps biologists study seal behavior as well as population dynamics,” Duper said. “Seals are an important indicator species for the ecosystem around them.”