A great tit recognised by the AI
Artificial intelligence has been trained to recognise individual birds, which is more than we humans are capable of. The system is being developed for biologists studying wild animals, but could be adapted so that people can identify individual birds in their surroundings.
André Ferreira at the Center for Functional and Evolutionary Ecology in Montpellier, France, started the project while studying how individual sociable weavers contribute to their colonies. This is normally done by putting coloured tags on their legs and sitting by nests to watch them, which is very time-consuming. Ferreira tried filming the colonies instead, but often the coloured tags weren’t visible in the footage, so he and his colleagues turned to AI.
The difficult part is getting the photographs required to train the system. “We need thousands of pictures of the same individual,” says Ferreira. “With humans, this is easy. With animals, it is hard to do.” The team solved this problem by putting RFID tags on the birds, which triggered cameras at bird feeders.
So far, the system has been tested on captive zebra finches, wild great tits and wild sociable weavers. Tests with photographs that weren’t used for training reveal its accuracy is around 90 per cent for a single image.
For now, the system is still quite limited. It has only been trained on pictures of the back of birds, as that is the view biologists usually get when observing behaviour. It might also fail if the appearance of a bird changes, for instance during moulting. “We don’t know exactly what the AI is using to identify birds,” says Ferreira.
However, he thinks all these issues can be overcome if given large-enough datasets. Ferreira and his team are now setting up cameras to take pictures from multiple angles, not just the back. The plan is to release the software for others to use as it is further developed.
There are already many AI-based apps available for identifying fauna and flora from images or sounds, but these identify only species, not individuals. Other teams are developing similar systems for identifying other animal individuals, says Ferreira, but this is the first that can identify small birds individually as far he knows. These systems could enable biologists to study individual behaviour in ways that are extremely difficult now, he says.
Journal reference: Methods in Ecology and Evolution, DOI: 10.1111/2041-210X.13436
More on these topics: