NOAA Facial Recognition Software May Help Save Right Whales
There may be a way to save the endangered population of North Atlantic right whales, scientists said.
North Atlantic right whales are baleen whales - the largest animals on Earth - that grow up to 50 feet and weigh up to 70 tons. Although they are enormous, right whales only feed on zooplankton, tiny species that wander about the ocean.
Right whales inhabit the Atlantic Ocean, and scientists said the animals' distribution depends on the distribution of their prey.
Unfortunately, the National Oceanic and Atmospheric Administration (NOAA) said there are only less than 500 individual right whales all over the world due to the effects of large-scale commercial fishing.
In April last year, the NOAA advised recreational boaters to keep their distance from a school of right whales to ensure the animals' safety, or else, boaters will be subjected to a hefty amount of fines. North Atlantic right whales are protected under the United States Endangered Species Act and the Marine Mammal Protection Act.
Now, NOAA scientists developed a facial recognition software for whales that could perhaps be a key in saving one of the planet's most endangered species.
How Facial Recognition May Save Right Whales
Researchers at the NOAA first set up a competition to better identify and distinguish whales.
The Right Whale Recognition contest, which was arranged by marine biologist Christin Khan, drew as many contestants as the current known population of right whales, with 470 contestants among 364 teams. Khan said she was inspired by Facebook's use of facial recognition to identify people in photos.
The winning entry uses a facial recognition algorithm to distinguish whales 87 percent of the time. The software identifies whales by the patterns on their heads. It makes use of artificial intelligence to align, localize and finally identify right whales from aerial photographs.
Developed by data science company deepsense.io, the facial recognition software could help save whales that have been caught in fishing nets. The algorithm allows scientists to report to disentanglement experts which whales have been trapped. At the same time, the algorithm would help marine biologists avoid performing mistaken biopsies on the same whale.
Most importantly, the software will save scientists countless hours spent trawling through images of whales, freeing up time to carry out actual research. Previously, the only source of locating and distinguishing whales was by using aerial survey flights. Cutbacks in funding have ended the flights in several places.
"Knowing how time-consuming post-flight processing can be, improved technologies to identify right whales quickly and accurately would be great progress," said research scientist Cynthia Taylor of the Sea to Shore Alliance.
Piotr Niedzwiedz, co-founder of deepsense.io, said it was very exciting for their Machine Learning Team to participate and then win the NOAA competition.
The solution they came up with helps solve a real-world problem and empowers marine biologists in their advocacies of protecting critically-endangered North Atlantic right whales, he said.
"“The Right Whale Recognition challenge was a great opportunity to put our data scientists’ talents to the test and to demonstrate that deep-learning techniques ... can provide immense benefits in big data applications,” said Niedzwiedz.
Meanwhile, Khan hopes to have the facial recognition software up and running by the next winter calving system.