Help ornithologists develop bird photo ID tool
Cornell lab’s computer identification of common North American avian species needs your photos
If you have a terrific picture of a Tennessee warbler, you can help the Cornell Lab of Ornithology improve its Merlin Bird Photo ID program. Upload your picture and put dots on the beak, eye and tail tip. Then, using patterns in the data, Merlin attempts to identify the bird. The aim is to help Cornell create a mobile device tool for beginning or intermediate birders to identify what they’ve snapped a picture of. (The lab already has a nice app that helps birders ID what they’ve seen based on a bird’s size, colors, behavior and location, but not with a photo.)
The tool is a fantastic idea but needs work. Developers are gradually adding species to the 400 common North American birds currently in the database. Each bird can be seen in many postures and plumages, so Merlin sometimes struggles to identify easy backyard birds. It gives users choices of possible species but little help in determining which is correct. An easy fix would be to link possible identifications to Cornell’s own All About Birds guide, which provides excellent information about each bird.