What is that bird? Bird Photo ID to the rescue
Ever spotted a bird next to the road and wondered what it is called? Without knowing its name, it’s challenging to find out more about the bird that caught your interest. Now you can ask Merlin!
Merlin, or Merlin Bird Photo ID, is a computer vision system that can identify hundreds of bird species. Currently in beta, Merlin can recognize 400 bird species commonly found in North America.
Using Merlin is very straightforward. (I will put firsthand user experience at the end.) First, you will upload a photo of a bird and tell it where and when the photo was taken. Then, you will be asked to draw a box around the bird and click on the bird’s beak, eye, and tail to help orient Merlin. After a couple of seconds (during which time Merlin uses its computer power to analyze the photo and compare it to millions of images and sightings recorded by birders), a short list of most likely matches are presented, along with photos, descriptions, and sound. How cool is that?
“[It’s] designed to keep improving the more people use it,” says Jessie Barry at Cornell Lab of Ornithology. So by using Merlin (which is free, by the way), you will be helping it become even more accurate.
“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,” says 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.”
Merlin is a fruit of the collaboration between researchers at Cornell Tech and California Institute of Technology and bird enthusiasts.
The following is my experience with Merlin (which I give thumbs up!). I had several photos of birds that I took a while back, clueless about their names.
I spotted this black seabird at a dock in San Francisco. It had very distinct facial features: orange skin connecting to the bill, and eyes set close to the edge of the orange skin.
Merlin identified the bird as Double-crested Cormorant, commonly seen on fresh and salt water across North America. It seems a perfect match.
Next, I tried a photo of a black bird with distinct red mark on its wings that I sighted on a trip to Monterey, California. I had seen these birds a number of times in Vancouver and the surrounding areas in Canada but had never known the name.
Merlin identified it as an adult male Red-winged Blackbird. (Well, this name shouldn’t be too hard to remember…)
Sometimes, birds are tricky to capture in photographs without decent cameras. When I unexpectedly spotted a bird, busy pecking at a moist ground and unfazed by the drizzle, I snapped a photo using my phone. The day was dark, and I was a few yards away — not the best condition for taking pictures with a phone.
I thought I would be pushing the limit with the low quality image, but Merlin still managed to identify the bird. Amazing! The bird is a Northern Flicker, a type of woodpecker. (More like a “ground-pecker”…)
Still, Merlin is not perfect. During a trip to Washington D.C., I spotted a blue bird from the bus, about to enjoy its wormy catch, when we stopped for the light. I only had time to snap a quick shot of it before the bus moved on.
Merlin’s best guess was Steller’s Jay. The shape looks very similar, but the coloring is wrong; Steller’s jay should have a dark head, but the bird in my photo has much lighter blue.
Having a pretty good idea that it is a jay, I searched for “blue jay bird” on a search engine and – Voila! – got the match, Blue Jay. Although Merlin got it wrong, it helped me narrow down my search.
Next time you spot a bird and want to learn more about it, why not ask Merlin? 🙂
Merlin is available at http://www.allaboutbirds.org/photoID. It’s also available as an app for smart phones. Enjoy!