Novel, Markerless, Kinematic Tool Utilising a Deep Neural Network for Analysis of Joint Range of Motion and Lameness In Dogs

Introduction: Kinematics is a widely accepted tool for gait assessment in dogs, however, its use requires specific equipment and knowledge which limits its clinical utility. Human research is now focused on markerless motion capture. Our study aimed to develop the first markerless kinematic tool to objectively measure the ROM of dogs’ joints. It also aims to identify changes associated with elbow osteoarthritis.

Materials and Methods: Thirty-eight large breed dogs were enrolled—5 sound control dogs and 33 dogs with elbow osteoarthritis. The dogs were filmed trotting with a GoPro4. The videos were uploaded to a computer-learning program which utilized a deep neural network to autonomously map the joints and spine. The range of motion (ROM) of the shoulder, elbow, hip, and stifle were calculated in the control dogs and compared with previously reported kinematic values to validate the program. The analysis was repeated on dogs with elbow OA to identify any changes in the data.

Results: The joint ROM in the control dogs were consistent with previously reported kinematic values. The program detected that elbow osteoarthritis in dogs caused a reduction in elbow ROM by 6.1 degrees.

Discussion/Conclusion: This novel markerless kinematic tool can accurately and precisely track joints of large breed dogs at the trot and provide joint ROM consistent with more invasive kinematic tools. The program identified a reduction in elbow ROM in dogs affected by elbow OA consistent with previous studies. This research is an important step towards objective, accessible, and noninvasive tools for lameness assessment in dogs.

Acknowledgment

None.

Publication History

Article published online:
15 July 2025