Scientists have created a website that can identify hundreds of birds through photographs.
The website AllAboutBirds.org/photoID was developed by the Visipedia research project in collaboration with the Cornell Lab of Ornithology, and is considered a breakthrough for bird watchers and researchers.
"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 at the Cornell Lab of Ornithology. "That's truly amazing, considering that the computer vision community started working on the challenge of bird identification only a few years ago."
The software, called Merlin Bird Photo ID, analyzes the pixels of a photo and combines artificial intelligence techniques with millions of data points from humans to determine the most likely species of bird in a matter of seconds. It has now been trained to recongnize hundreds of bird species.
"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. "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."
The computer learns to recognize each species from tens of thousands of photographs labeled by ornithologists and draws from over 70 million sightings recorded in the eBird.org database. It also narrows down possible species by the location and time of year the bird was spotted.
The development will be presented at Cornell Tech and the California Institute of Technology at the Computer Vision and Pattern Recognition (CVPR) conference.
