Researchers studying sharks are able to identify individual sharks by the markings on the edge of their dorsal fins. Much like a fingerprint, each shark has its own unique pattern of bumps, notches and scars. When studying shark populations, researchers have had to manually compare old images with new ones to sort out sharks they've already identified from new ones, a task that can be very time consuming.
Dr. Sara Andreotti, a marine biologist in the Department of Botany and Zoology at Stellenbosch University in South Africa knew there had to be a better way. For six years, she had built a database of great white sharks she's seen off the coast of South Africa, with profiles on each individual, including DNA information if she and her colleagues had been able to collect a biopsy. She wanted to have a faster way to pair new photographs with her detailed database.
Andreotti sought help from the university's applied math department where a specialist in machine learning knew exactly how to tackle the problem. They built an image recognition software called Identifin that traces a line along the notches on the back edge of the dorsal fin in a photograph and then matches that line to existing images in the database. The existing images are ranked in order of likelihood of it being a match, with the photo in the number one spot being the correct one if it's an already known shark.
© Stellenbosch University
If the photo in the number one spot doesn't match, it's a new shark.
"Previously, while at sea, I had to try and memorize which shark is which, to prevent sampling the same individual more than once," said Andreotti. "Now Identifin can take over. I will only need to download the new photographic identifications from my camera onto a small field laptop and run the software to see if the sharks currently around the boat have been sampled or not."
"By knowing which sharks had not been sampled before we can focus the biopsy collections on them. This saves us both time and money when it comes to genetic analysis in the laboratory."
On a larger scale, if software like this can be become the industry standard for marine biologists, researchers will be able to compare their data to others around the world and get a full picture of the distribution great white sharks and other species as well.
The next step for the team is to tweak the software so that it can be used for a variety of large marine animals and make it accessible to other researchers.
Our editors found this article on this site using Google and regenerated it for our readers.