Automatic Detectors Of Animal Behaviors Using Computers
Many millions of people all over the world live with companion animals; as technology progresses and the regular home becomes the “smart home” of the future, this shift will inevitably have an impact on the lives of our companion animals. Many of us have seen videos of dogs and cats “using” smartphones and tablets in their own ways, but such technologies are typically designed with humans in mind. An emerging field, dubbed “Animal-Computer Interaction (or ACI),” is taking a different approach, by targeting animals as ” users of digital systems.” Within this field, “computer interfaces and digital systems” are developed specifically for animal use. ACI has the potential to not only revolutionize the way companion animals’ lives are enriched within our homes, but in other places as well, by fostering physical activity in shelters or providing enrichment in captivity such as zoos and labs.
The development of this field relies on advancements in computer vision and machine learning that would automate behavior recognition in animals and react accordingly. Since most non-human animals are predominantly non-verbal communicators, the programming must incorporate an understanding of the body postures and patterns, as well as the behaviors and moods these actions indicate. Whether this programming takes the form of wearable devices or the use of cameras, these behaviors need to be quickly and accurately recognized by computers so that they can react appropriately.
This particular research project describes the development of a “non-wearable depth-based tracking system for cats,” and its promising results. The testing was done with cats and their human companions, in order to contribute to ACI research. Researchers discovered that a tracking system based on depth information is possible, and also that supervised and “knowledge-based” approaches both worked well. The results show the potential benefits of a tracking system that could “effectively detect a cat’s location and also use classification algorithms to recognize a cat’s postures and body parts,” and make changes to their environments to suit their enrichment needs.
Companion animal advocates—or advocates for animals in captivity—may balk at the idea of computers enriching the lives of companion animals or captive animals, but these types of technologies may be somewhat of a forgone conclusion. As humans continue to accept and incorporate computers into their day-to-day environment, it seems probable that non-human animals will be affected by this shift. It will be up to advocates to raise their voices to try to steer such technologies in a positive direction in order to benefit as many animals as possible.