Buzz
Beyond social media tagging, facial recognition technology is now being utilized to distinguish individual whales from aerial images.
Christin Khan, a biologist with the National Oceanic and Atmospheric Administration, focuses on North Atlantic right whales. She noticed unique pale markings, known as callosities, on the whales' heads. These raised skin patches are covered in thousands of tiny crustaceans called cyamids, often referred to as "white lice," giving them a light hue. Like human fingerprints, these patterns are unique to each whale, allowing researchers to tell them apart.
Khan transformed her idea into a research tool by launching a competition on Kaggle, offering a $10,000 prize sponsored by MathWorks. Nearly 500 participants competed, with the data science firm Deepsense.io winning for their software, which achieved 87% accuracy in identifying whales. The project began with 4500 whale images, and the team aims to gather more photos to enhance the technology further.
Beyond aiding whale research, this software holds potential for direct conservation efforts. For instance, if a whale gets caught in fishing nets, crews could use the technology to identify the animal and access its health records, helping them decide the best rescue approach. Khan estimates it will take another year before the software is field-ready. Meanwhile, similar identification methods are being developed for humpback whales, which are distinguished by their unique tail flukes rather than back markings.
