Birdsnap: Bird Nerding on a Smartphone
WHY YOU SHOULD CARE
Because now you don’t have to read all the wildlife identification signs in the nature park. Just turn to your phone.
By Vignesh Ramachandran
Bird enthusiasts have something new to be enthusiastic about: An app that can spot a bird species with a simple picture.
Researchers at Columbia University and the University of Maryland are using computer vision and machine learning to automatically generate a field guide for bird species.
Goodbye, paper book. Hello, new app.
If you’ve ever been bird-watching, part of the experience (indeed the joy of it) is figuring out exactly what the heck you are looking at. The researchers’ new app (free for iOS ) and website — Birdsnap — uses a super smart algorithm to identity parts of a bird and then attempt to pinpoint its species.
All you have to do is take a photo, upload it via the app or website to Birdsnap servers, identify the head and tail of the bird and the technology does its magic. The bird’s body parts — eye, beak, neck, wing or feet — are analyzed. The electronic field guide is focused on 500 of the most common bird species in North America, so it’s not currently catered to birdwatchers across the world.
Birdsnap also has a teaching component: the program shows you how to identify key parts of a bird, so your brain can be its own field guide someday. The website is interactive with descriptions and lets you hear the distinct sounds of each species.
Belhumeur previously helped develop Leafsnap, electronic field guides that identifies tree species from photos of leaves. Before you wonder who would actually use such a thing, note that Leafsnap has more than a million users.“It’s fun to come in as a computer scientist to say, ‘Here’s something we can do with your (biologists’) data,’” Peter Belhumeur, a Columbia computer science professor, said in a statement. “I get to take this technology and move it into a space that I am very interested in.”All this data also allows researchers to pinpoint where a particular species might be located across the United States on a particular day. Birdsnap shows a map that displays the likelihood of sighting throughout the year. In the rapidly growing world of big data, Birdsnap is one example of how biology and machine learning can be married.
We’ve seen similar face recognition technology used in consumer photo apps or sites that many of us regularly use for humans. Facebook is smart enough to suggest people you should tag in your photos. Apple iPhoto’s “Faces” feature scans people’s faces in your photo library so you can better organize your collection.
With emerging wearables like Google Glass, some uses of facial-recognition technology could even match strangers with their online profiles. Can you imagine the day when you’re on the airplane and you could immediately know your seatmate’s life story via his or her digital footprint — simply by looking at them? Sounds like a future alluded to in 2002’s Minority Report.
But before we go all everyone-is-Big-Brother-with-their-creepy-wearables, you can spend the summer getting up close and personal with your avian friends outside. Just don’t get whitewashed when you lean in for an iPhone snap.