Is Markerless More or Less? Comparing a Smartphone Computer Vision Method for Equine Lameness Assessment to Multi-Camera Motion Capture.

Computer vision is a subcategory of artificial intelligence focused on extraction of information from images and video. It provides a compelling new means for objective orthopaedic gait assessment in horses using accessible hardware, such as a smartphone, for markerless motion analysis. This study aimed to explore the lameness assessment capacity of a smartphone single camera (SC) markerless computer vision application by comparing measurements of the vertical motion of the head and pelvis to an optical motion capture multi-camera (MC) system using skin attached reflective markers. Twenty-five horses were recorded with a smartphone (60 Hz) and a 13 camera MC-system (200 Hz) while trotting two times back and forth on a 30 m runway. The smartphone video was processed using artificial neural networks detecting the horse's direction, action and motion of body segments. After filtering, the vertical displacement curves from the head and pelvis were synchronised between systems using cross-correlation. This rendered 655 and 404 matching stride segmented curves for the head and pelvis respectively. From the stride segmented vertical displacement signals, differences between the two minima (MinDiff) and the two maxima (MaxDiff) respectively per stride were compared between the systems. Trial mean difference between systems was 2.2 mm (range 0.0-8.7 mm) for head and 2.2 mm (range 0.0-6.5 mm) for pelvis. Within-trial standard deviations ranged between 3.1-28.1 mm for MC and between 3.6-26.2 mm for SC. The ease of use and good agreement with MC indicate that the SC application is a promising tool for detecting clinically relevant levels of asymmetry in horses, enabling frequent and convenient gait monitoring over time.

Visual lameness assessment in comparison to quantitative gait analysis data in horses.

BACKGROUND: Quantitative gait analysis offers objective information to support clinical decision-making during lameness workups including advantages in terms of documentation, communication, education, and avoidance of expectation bias. Nevertheless, hardly any data exist comparing outcome of subjective scoring with the output of objective gait analysis systems. OBJECTIVES: To investigate between- and within-veterinarian agreement on primary lame limb and lameness grade, and to determine relationships between subjective lameness grade and quantitative data, focusing on differences between (1) veterinarians, (2) live vs video assessment, (3) baseline assessment vs assessment following diagnostic analgesia. STUDY DESIGN: Clinical observational study. METHODS: Kinematic data were compared to subjective lameness assessment by clinicians with ≥8 years of orthopaedic experience. Subjective assessments and kinematic data for baseline trot-ups and response to 48 diagnostic analgesia interventions in 23 cases were included. Between and within-veterinarian agreement was investigated using Cohen's Kappa (κ). Asymmetry parameters for kinematic data ('forelimb lame pattern', 'hindlimb lame pattern', 'overall symmetry', 'vector sum head', 'pelvic sum') were determined, and used as outcome variables in mixed models; explanatory variables were subjective lameness grade and its interaction with (1) veterinarian, (2) live or video evaluation and (3) baseline or diagnostic analgesia assessment. RESULTS: Agreement on lame limb between live and video assessment was 'good' between and within veterinarians (median κ = 0.64 and κ = 0.53). There was a positive correlation between subjective scoring and measured asymmetry. The relationship between lameness grade and objective asymmetry differed slightly between (1) veterinarians (for all combined parameters, p-values between P < .001 and 0.04), (2) between live and video assessments ('forelimb lame pattern', 'overall symmetry', both P ≤ .001), and (3) between baseline and diagnostic analgesia assessment (all combined parameters, between P < .001 and .007). MAIN LIMITATIONS: Limited number of veterinarians (n = 4) and cases (n = 23), only straight-line soft surface data, different number of subjective assessments live vs from video. CONCLUSIONS: Overall, between- and within-veterinarian agreement on lame limb was 'good', whereas agreement on lameness grade was 'acceptable' to 'poor'. Quantitative data and subjective assessments correlated well, with minor though significant differences in the number of millimetres, equivalent to one lameness grade between veterinarians, and between assessment conditions. Differences between baseline assessment vs assessment following diagnostic analgesia suggest that addition of objective data can be beneficial to reduce expectation bias. The small differences between live and video assessments support the use of high-quality videos for documentation, communication, and education, thus, complementing objective gait analysis data.

Kinematic and spatiotemporal assessment of habituation to treadmill walking in Labrador retrievers.

BACKGROUND: This study investigated differences in kinematic and spatiotemporal variables in Labrador retrievers during introduction to treadmill walking, with the aim to determine the time required for them to become habituated. Twenty-five healthy, treadmill-naive Labrador retrievers participated in the study. The total angular displacement of the carpus, elbow, tarsus and stifle, as well as stride time and stance duration were calculated from the 3-D tracking of skin mounted reflective markers recorded with 6 infrared light emitting video cameras at 240 Hz. The measurements were done at two walking speeds, 0.78 and 0.96 m/s, in six sessions on the treadmill during two consecutive days. RESULTS: With a 1-2 min acclimatization period following each treadmill speed change, mean values of the study variables were significantly different from the last training session mainly in the first session on the first day. However, between-stride variability was significantly larger for at least one variable even in the fourth session for the slower walking speed, and in the fifth session for the higher walking speed. CONCLUSIONS: The results show the importance of proper pre-training of dogs in locomotion studies at walk using a treadmill, and the need to consider not only variable mean values but also between-stride variability, in order to ensure that dogs are sufficiently accustomed to allow collection of reliable data.