People generally associate facial recognition with surveillance, Apple’s Face ID, and perhaps the biometric payment methods being deployed in China. The conservation of seals probably doesn’t make the list. The technology has found use for this exact application, as unlikely as this may seem, and it could see broader adoption. At Colgate University, a research team developed software called SealNet, an artificial intelligence tool designed to aid in the long-term monitoring of harbor seals. Traditionally, the seals have been monitored using GPS tags and trackers, which are costly and can interfere with the animal’s natural behavior. Some progress has been made with photo-identification techniques, but these methods have proven unreliable because identification is based on spots and patterns in a seal’s coat, which change over time. Courtesy of Pixabay/Susanne Jutzeler. Encouraged by success with facial recognition technology developed for other mammals — such as lemurs, brown bears, and pandas — the team developed SealNet, which is based on similar software developed to identify primates. The SealNet technology was trained on a database of seal faces created from more than 1700 images of over 400 individual harbor seals in Maine’s Casco Bay. As with facial recognition for humans, SealNet is designed to automatically detect the face in an image and compare it against facial patterns in its database. The team tested the software against PrimNet, which was retrained on seals. SealNet significantly outperformed its predecessor. According to the researchers, the technology showed that it could identify seals with 96% accuracy — no small feat in an ecosystem home to thousands of the marine animals. The team is working to expand its database to make it available to other scientists. Broadening the database to include species such as Mediterranean and Hawaiian monk seals could help inform conservation efforts to save these rare species. The technology could also help scientists to understand where in the ocean seals are traveling. The Colgate team published its findings in April in the scientific journal Ecology and Evolution. The processed images were taken across seven locations during the summers of 2019 and 2020.