A digital transformation of orthopaedic disease detection in large animals using artificial intelligence

Senast ändrad: 04 maj 2023

Elin Hernlund.

Starting from the research history in the field of horse locomotion, this lecture will cover how technological development can be used for successful improvements of the diagnostic efficiency in veterinary medicine. Horse lameness assessment is today increasingly performed by human vision and computer vision in unison, which leads to improved clinical decision-making. The use of artificial intelligence for medical purposes carries a huge potential, if used wisely.

Eyesight is the richest of the human senses. By watching an animal’s body shape a human with animal experience can figure out information about species, breed, gender, age, conformation, body condition score (nutrition), and training status. From the body pose we can tell if the animal is alert or resting, and from the movement we can see behavioural display as well as signs of disease such as lameness or colic.

Today, a computer could do the same thing. It’s “eyesight” could be developed to classify high context information from videos or images. This technological field is called computer vision and is a sub field of artificial intelligence. Computer vision provides means for great progress in disease detection in human medicine. The computer can be trained to correctly detect breast cancer in X-ray images from mammography screening, metastases in tissue sections and to classify skin cancer from photos of moles. In fact, the results show that the computer can outperform the detection capacity and performance of trained medical specialists due to its pattern recognition ability and thus the potential to efficiently make sense of complex data.

Implementation of technological innovations for medical diagnostic purposes has a long and interesting history. There are many examples of fierce resistance from the medical community against the use of technical devices. In hindsight, this scepticism can appear surprising. As an example, the use of thermometers to measure body temperature was not accepted by many physicians even though body temperature was used as an important diagnostic sign from the early days of medicine.

Researchers in the field of animal locomotion were very early to adapt technology to aid human perception and understanding of the complex motion of four-legged animals. This curious use of technology is a good example of how interdisciplinary research can lead to important innovations that can be used in medical clinical practice. Starting in the 1870ies, photographic methods revolutionized our understanding of horse gaits. In the 21st century, animal locomotion studies have been heavily driven by the increasing access to motion sensors and other motion capture techniques. Today, AI methods, such as computer vision are explored to improve disease detection capacity through of animal body motion.