Researchers from the Tech4Animals lab at the University, São Paulo University in Brazil, Lincoln University, and Nottingham University in the UK developed groundbreaking artificial intelligence capable of identifying pain in cats.
Detecting pain in cats is a challenging task even for vets due to their evolutionary ability to hide discomfort. House cats often suffer from chronic pain, but their humans find it difficult to tell when they are in pain and consequently do not take them to the vet.
But now, with the use of deep learning models and facial recognition analysis, the researchers have achieved an impressive success rate of over 70% in identifying which cats are in pain. The methods were tested on 29 Short Hair British cats, and the areas around the mouth and eyes were found to be most significant in identifying pain.
The international study involving researchers from the University that was published in the prestigious journal Scientific Reports from the Nature group was led by doctoral student Marcelo Feighelstein with his advisors Prof. Anna Zamansky, and Prof. Ilan Shimshoni from the Department of Information Systems at the University.
The team photographed the female cats’ faces before and after sterilization, while they are still under the influence of painkillers, and after the painkillers have worn off. Models detected the subtle changes in facial expressions such as the tips of the ears, eyes, whisker and mustache that indicate pain.
This incredible breakthrough has the potential to revolutionize the way we care for our feline friends, making it easier for anyone who has a cat to photograph them, without the need for physical contact, and to check reliably whether the cat is in pain.
Tech4Animals Lab is working on a digital Dr. Dolittl-E app to detect animal feelings is showing how AI technology is advancing and transforming the world of veterinary care, and how it could help us better understand our animal companions.