Computer system scans video footage to analyze animal behavior

Usually, when studying animal behavior, scientists have to watch many hours of video footage of the creatures going about their daily lives. That may soon no longer be necessary, however, thanks to a new creature-watching computer system.

Currently being developed by a team from the University of Zurich and the ETH Zurich research institute, the technology uses an image analysis algorithm that uses computer vision and machine learning.

When used to analyze long video recordings of animals in their natural habitat or in captivity, it can differentiate between individual animals and identify behaviors such as those associated with fear, curiosity, or harmonious social interactions with other members of the same species.

Not only should the system save researchers from having to spend days or even weeks looking at these images themselves, but it should also help standardize the analysis of recordings, which might otherwise be subject to bias by scientists. individual. Crucially, it’s additionally capable of detecting behavioral changes that occur gradually over time, which might be too subtle to notice by humans reviewing hours and hours of footage.

The system was trained on videos of mice and macaque monkeys in captivity, but should be applicable to all animal species. It is already being tested at Zurich Zoo and in a study with wild chimpanzees in Uganda.

The technology could also be used to improve animal welfare in captivity, by identifying behaviors associated with underlying issues earlier than would otherwise be possible. In the case of animals used in laboratory studies, the system could potentially detect signs of stress or discomfort earlier, thereby minimizing the amount of suffering involved.

An article on the research, led by Markus Marks, a postdoctoral student at ETH Zurich, was recently published in the journal Intelligence of natural machines.

Source: ETH Zürich

Gordon K. Morehouse