What shines here in fluorescent pink, yellow, green and blue could be the future of behavioral observation.
The precise observation of animal and human behavior is essential for many areas of research – especially in biology. They allow scientists to find out, for example, how animals interact with each other, whether species protection measures work and how populations develop. Over time, behavioral research methods have continued to improve. Scientists no longer have to rely on pens and clipboards to record an animal’s movements, but can increasingly rely on digital helpers.
A major technological leap was the possibility of attaching special markers to animals, which are then recorded by a computer program. But some animals resisted the markers or were simply too small to be equipped with them. However, it is now possible to research an animal’s behavior without any markers – thanks to adaptive algorithms. To do this, scientists simply have to teach the special program to recognize the animals and their postures in drone images or videos. However, training is very laborious and requires extensive manual work, and these systems often fail as soon as the camera or light settings change. Thus, although behavioral observation has become more automated, it still relies heavily on human help.
Researchers led by Daniel Butler from the Salk Institute for Biological Studies in California may now have developed a technology that solves the previous problems. The idea: You paint relevant body parts of the animals being observed with fluorescent paint, which is stimulated to glow by wavelengths outside the visible range. By then creating pairs of images in which the color is visible and invisible, the AI is trained to recognize the body parts by itself at some point, even if there is no color attached to the test animals or people.
Once you have captured a rabbit with the system in the laboratory, you can theoretically show the digital assistant video footage of wild rabbits and it will then transfer its skills to it without you having to retrain it first. The new technology is called “GlowTrack” and its inventors expect great things from it. “Our approach can benefit a variety of areas that need more sensitive, reliable and comprehensive tools to detect and quantify movement,” says Butler’s colleague Eiman Azim. “I’m excited to see how other scientists and non-scientists adopt these methods and what unique, unforeseen applications they might lead to.”