Ever looked at a bird and been completely stumped when trying to figure out what it is? No need to ruffle your feathers bird lovers, as a new web-based tool is flying to the rescue. The newly developed ‘Merlin Bird Photo ID’ acts as your own personal ornithologist. So how does it work? All you need to do is take a snap of a bird you're keen to identify, upload it to the website and let Merlin do the rest.
The website, brought to you by the Cornell Lab of Ornithology and the Visipedia research project, can recognize 400 of the most commonly found birds in the United States and Canada. Once users have taken a photo, they draw a box around the bird, click its bill, eye and tail and tell Merlin where and when the photo was taken.
“It gets the bird right in the top three results about 90 percent of the time, and it’s designed to keep improving the more people use it,” said Jessie Barry, the Merlin Project Leader, in a statement. “That’s truly amazing, considering that the computer vision community started working on the challenge of bird identification only a few years ago.”
Though the program was named Merlin due to its “uncanny” and “magical” ability to recognize hundreds of birds, the real magic comes from the invaluable contribution from bird lovers who have taken and labeled thousands of images. Merlin uses the eBird.org database, which contains more than 70 million bird sightings, to help identify specific bird species. Within a few seconds, Merlin presents you with the top three best matches to your image, with a range of photos and sounds. The website is free to use, and researchers are currently working on a mobile phone app.
“Computers can process images much more efficiently than humans – they can organize, index and match vast constellations of visual information such as the colors of the feathers and shapes of the bill,” said Serge Belongie, a professor of computer science at Cornell Tech, in a statement. “The state-of-the-art in computer vision is rapidly approaching that of human perception, and with a little help from the user, we can close the remaining gap and deliver a surprisingly accurate solution.”
