Is Markerless More or Less? Comparing a Smartphone Computer Vision Method for Equine Lameness Assessment to...

Area of Science:

• Veterinary Medicine
• Biomechanics
• Computer Vision

Background:

• Objective lameness assessment in horses is crucial for early detection and management of musculoskeletal conditions.
• Traditional methods often require specialized equipment and controlled environments.
• Markerless motion analysis using computer vision presents a more accessible approach.

Purpose of the Study:

• To evaluate the lameness assessment capabilities of a smartphone single camera (SC) markerless computer vision application.
• To compare vertical motion measurements of the horse's head and pelvis obtained via SC with a multi-camera (MC) optical motion capture system.

Main Methods:

• Twenty-five horses were recorded trotting on a runway using both an SC (60 Hz) and an MC (200 Hz) system.
• Artificial neural networks processed smartphone videos for motion analysis.
• Vertical displacement data from head and pelvis were synchronized and compared between systems.

Main Results:

• The SC system demonstrated a mean difference of 2.2 mm for both head and pelvis vertical motion compared to the MC system.
• Within-trial standard deviations showed good agreement between the SC and MC systems.
• The SC application exhibited minimal differences in measurements, indicating high precision.

Conclusions:

• The smartphone computer vision application is a promising tool for objective orthopaedic gait assessment in horses.
• Its ease of use and good agreement with established systems enable frequent and convenient gait monitoring.
• This technology can facilitate the detection of clinically relevant gait asymmetries over time.