A research team from the University of Florida College of Veterinary Medicine has managed to use artificial intelligence (AI) to identify pain in goats based on their faces. The researchers think the system can also be applied to other animals and humans that cannot communicate verbally.
Advertisement
The assessment of whether an animal is in pain or not is often subjective and based on years of experience. However, not every animal can see or hear equally well whether it feels pain or not. The researchers wanted to find out whether there are clear features in the faces of goats that indicate pain. They published their findings in the study “Automated acute pain prediction in domestic goats using deep learning-based models on video recordings” in Scientific Reports.
The scientists created an AI model that was trained with videos of the faces of goats that were both in pain and not in pain. The model was trained with videos of a total of 40 goats. The researchers then tested the system on several goats. The system achieved recognition performance of up to 80 percent in identifying pain goats.
Can also be applied to other animal species and humans
The researchers assume that their AI system can also be applied to other animal species, where it is difficult to recognize whether they are experiencing pain. With the right training, the AI model could also be applied to them, and in general to all patients who cannot communicate verbally – including certain people.
“If we solve the problem in animals, we can solve the problem in children and other nonverbal patients,” said Ludovica Chiavaccini, professor of anesthesiology at the University of Florida College of Veterinary Medicine.
Based on the research results, the researchers also developed a pain scale for goats. Such a scale did not previously exist. However, the established scale cannot be generalized at this time because the underlying data was only collected for male goats during castration.
(olb)
Don’t miss any news – follow us on Facebook, LinkedIn or Mastodon.
This article was originally published in German. It was translated with technical assistance and editorially reviewed before publication.
