Scientists have created a neural network that distinguishes seals by their faces.

SealNet developers trained a neural network to distinguish between harbor seals using photographs of 406 different individuals. They then tested the software and found that SealNet could correctly identify the animal 85% of the time based on distinctive facial features. According to scientists, this is necessary to track the movements of seals. The team's tests showed that some seals return to the same places year after year. However, other seals appeared in other places. An accurate understanding of how seals move can help to more accurately identify areas of their habitat that need to be protected from poachers.