Even the most talented birdwatcher can have difficulty in identifying individual birds within the same species. Now researchers have developed an artificial intelligence tool which can do the job for them. The Times journalist, Tom Knowles, reports that this could have a radical effect on the study of birds because it will allow more detailed and reliable monitoring and assessment going forward.
Reliably recognising individual birds is said to be one of the greatest limitations in their study according to André Ferreira of the Centre for Functional and Evolutionary Ecology in France.
French, German, Portuguese and South African scientists showed the artificial intelligence models sounds and images of wild great tits, sociable weavers and captive zebra finches. The software has an accuracy of 90% for the wild birds and for zebra finches it had an 87% success rate when shown images that it had not seen before.
The study is published in Methods in Ecology and Evolution. I visited their website and using their search facility could not find the study. It may not have been published at this time and The Times journalist may have seen it before being aired on their website.
