- Title: Can AI read the emotions in animal faces?
- Date: 18th March 2025
- Summary: BRISTOL, ENGLAND, UK (MARCH 4, 2025) (Reuters) (SOUNDBITE) (English) MELVYN SMITH, PROFESSOR OF MACHINE VISION & DIRECTOR OF THE CENTRE FOR MACHINE VISION (CMV), UNIVERSITY OF THE WEST OF ENGLAND BRISTOL, SAYING: "So the idea is that the system would be integrated with the automated feeder, so there would be a camera built into the feeder and some lighting looking at t
- Embargoed:
- Keywords: AI UWE Bristol University of the West of the England Bristol animal emotions animals artificial intelligence computer vision farming machine learning
- Location: BRISTOL, ENGLAND, UK / UNSPECIFIED LOCATION / VARIOUS FILE LOCATIONS
- City: BRISTOL, ENGLAND, UK / UNSPECIFIED LOCATION / VARIOUS FILE LOCATIONS
- Country: UK
- Topics: Europe,Life Sciences,Science
- Reuters ID: LVA004827512032025RP1
- Aspect Ratio: 16:9
- Story Text: Can you read the emotions of animals using artificial intelligence (AI) just by looking at their faces? That's the hope of Melvyn Smith, a machine vision engineer at the University of the West of England Bristol (UWE) who is developing a system to be integrated into an automated feeder.
The IntelliPig, developed by UWE Bristol and Scotland’s Rural College (SRUC), utilizes a camera installed within the feeder, where the AI captures the facial features of the feeding pig, enabling the system to recognize the individual animal, assess its condition, and detect signs of stress.
"The beauty of that is that when the pig comes to drink, you can get a good view of its face by locating the camera in the right place. So the concept here really is to see what's the maximum we can get out of something like that," Smith told Reuters.
Stress is detected by a deep learning technique whereby the AI is trained by showing it lots of examples of animals that are stressed and lots that aren't stressed, allowing it to discern subtle traits of stress in a pig's face by comparing images of the animal in different emotional states. The system could not just help to identify the presence of stress, but measure its level, Smith added.
However, the goal extends beyond merely identifying stress. Smith said their ambition is to "expand the spectrum from the negative into the positive", which involves shifting from a binary understanding of animal emotions to measuring the levels of stress and interpreting a broader range of emotions.
"AI lends itself very well to that. There are techniques that can classify and there are also techniques that can measure. So we want to be looking at the latter to try and plot the animal in this emotional space," he said.
While pigs are the target of the current research, it could be easily transferred to other animals, Smith said.
The ultimate aim of the IntelliPig system is not just to ensure that animals live a stress-free life, but also a happy one. This could potentially lead to significant advances in animal welfare, ushering in a new era of caring for animals that prioritizes their health, happiness, and well-being. Future developments may even allow the AI to interpret more complex emotions akin to human experiences of joy, serenity, frustration, or fear, further enhancing our understanding of animal emotions and welfare.
The technology is being designed with user-friendliness in mind, particularly for farmers. Smith envisions a future where the system presents data through a mobile app or laptop interface, enabling farmers to monitor their animals more effectively.
(Production: Matt Stock) - Copyright Holder: REUTERS
- Copyright Notice: (c) Copyright Thomson Reuters 2025. Open For Restrictions - http://about.reuters.com/fulllegal.asp
- Usage Terms/Restrictions: None