started · updated
AI frameworks advance animal welfare and pain detection
Researchers are advancing the use of artificial intelligence to monitor the well-being of animals and patients who cannot communicate pain or behavior.
A team led by Dr. Marcelo Feighelstein at Tel-Hai University has developed the SHIC-XE framework. This system uses video to detect pain in horses by mapping 2D images to a 3D anatomical model. Unlike previous “black box” AI models that relied on unstable heat maps, SHIC-XE provides explainable results by focusing on consistent anatomical coordinates, such as ear position or muscle tension. The system's findings align closely with veterinary experts and show potential for future application in monitoring human patients in intensive care or those with dementia.
In Taiwan, the National Health Research Institutes and Tsing Hua University have developed an AI imaging system designed to make animal experimentation more humane. This system acts as a 24-hour caregiver, capable of identifying at least eight distinct behaviors—including eating, drinking, and sleeping—with an overall accuracy exceeding 80%. By converting qualitative observations into quantitative data, the technology supports the ‘3R principles’ of animal research (Replacement, Reduction, and Refinement), allowing researchers to detect physiological or behavioral changes more precisely and early.
Entities
Chen Ren-kun · Marcelo Feighelstein · National Health Research Institutes · National Tsing Hua University · Tel-Hai University