In the UK, researchers have developed a system that analyzes photos of pigs. It's called Intellipig and helps pig farmers determine if their animals show signs of pain, illness, or stress.
Similar research is underway elsewhere in the world, involving, among other things, sheep, horses, and cats, reports Science. By photographing horses' faces before and after operations or before and after receiving pain relief, AI can learn to recognize when an animal signals pain or not.
That animals change facial expressions with different emotions is nothing new, but while it can take long training and several minutes for a human to assess whether a horse is in pain or not, AI can do it in a fraction of the time.
Horses have many signals, such as how their ears are angled and small muscle tensions around the mouth and eyes. In one study, AI managed to determine whether a horse was in pain or not in 88 percent of cases, significantly faster and better than humans. In another study with researchers from, among others, the University of Haifa, AI outperformed trained veterinarians in detecting pain in goats.
The same research team has also trained AI to determine with 77 percent accuracy whether a cat is in pain.
Perhaps AI can help us read our pets' faces. But the researchers also hope that facial recognition can become a way to promote animal health and welfare by quickly identifying animals that are doing poorly.
