A simple method for equine kinematic gait event detection

Holt, Danielle; George, Lindsay Blair St; Clayton, HM; Hobbs, Sarah Jane

doi:10.1111/evj.12669

Cited by 9 publications

(9 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nevertheless, the difference was marginal and there is great advantage in using models based on raw sensor data. It is extremely challenging and time consuming to develop specific algorithms for feature extraction [14][15][16] . These algorithms require validation, and they risk being gait, surface and ultimately, breed specific.…”

Section: Discussionmentioning

confidence: 99%

Improving gait classification in horses by using inertial measurement unit (IMU) generated data and machine learning

Bragança

Broomé

Rhodin

et al. 2020

Sci Rep

View full text Add to dashboard Cite

For centuries humans have been fascinated by the natural beauty of horses in motion and their different gaits. Gait classification (GC) is commonly performed through visual assessment and reliable, automated methods for real-time objective GC in horses are warranted. In this study, we used a full body network of wireless, high sampling-rate sensors combined with machine learning to fully automatically classify gait. Using data from 120 horses of four different domestic breeds, equipped with seven motion sensors, we included 7576 strides from eight different gaits. GC was trained using several machine-learning approaches, both from feature-extracted data and from raw sensor data. Our best GC model achieved 97% accuracy. Our technique facilitated accurate, GC that enables in-depth biomechanical studies and allows for highly accurate phenotyping of gait for genetic research and breeding. Our approach lends itself for potential use in other quadrupedal species without the need for developing gait/animal specific algorithms.

show abstract

Section: Discussionmentioning

confidence: 99%

Improving gait classification in horses by using inertial measurement unit (IMU) generated data and machine learning

Bragança

Broomé

Rhodin

et al. 2020

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Kinematic data were interpolated and low-pass filtered (Butterworth 4 th order), with a cut-off frequency of 12 Hz, as determined using residual analysis. FL and HL hoof impact and lift-off events were calculated from kinematic data in accordance with Holt et al (2017), using the FL and HL sagittal plane angles illustrated in Figure 1. Angle-time curves were plotted for FL and HL sagittal plane angles.…”

Section: Data Processing and Analysismentioning

confidence: 99%

“…Angle-time curves were plotted for FL and HL sagittal plane angles. An angle of zero degress was observed when the two segments were aligned and was used as a threshold for gait event detection for the FL and HL, with hoof impact and lift-off coinciding with descent and ascent and through zero degrees, respecitively (Holt et al, 2017).…”

Section: Data Processing and Analysismentioning

confidence: 99%

The effect of cut-off frequency when high-pass filtering equine sEMG signals during locomotion

George

Hobbs

Richards

et al. 2018

Journal of Electromyography and Kinesiology

Self Cite

View full text Add to dashboard Cite

High-pass filtering (HPF) is a fundamental signal processing method for the attenuation of low-frequency noise contamination, namely baseline noise and movement artefact noise, in human surface electromyography (sEMG) research. Despite this, HPF is largely overlooked in equine sEMG research, with many studies not applying, or failing to describe, the application of HPF. An optimal HPF cut-off frequency maximally attenuates noise while minimally affecting sEMG signal power, but this has not been investigated for equine sEMG signals. The aim of this study was to determine the optimal cut-off frequency for attenuation of low-frequency noise in sEMG signals from the Triceps Brachii and Biceps Femoris of 20 horses during trot and canter. sEMG signals were HPF with cut-off frequencies ranging from 0 to 80 Hz and were subjected to power spectral analysis and enveloped using RMS to calculate spectral peaks, indicative of motion artefact, and signal loss, respectively. Processed signals consistently revealed a low-frequency peak between 0 and 20 Hz, which was associated with motion artefact. Across all muscles and gaits, a 30-40 Hz cut-off fully attenuated the low-frequency peak with the least amount of signal loss and was therefore considered optimal for attenuating low-frequency noise from the sEMG signals explored in this study.

show abstract

“…Values generated from sound individuals can serve as a baseline to compare with clinical cases. A study by Holt et al [11] proposed a simple kinematic method using force plates for detecting hoofon, hoof-off, and peak ground reaction forces at walk, trot, and canter with acceptable accuracy. Another study evaluated stride detection using inertial measurement units (IMUs) [12].…”

mentioning

confidence: 99%

Practical uses of quantitative gait analysis in horses

Álvarez

Weeren

2019

Equine Veterinary Journal

View full text Add to dashboard Cite

When we hear 'gait analysis' we immediately think 'lameness'. This is not surprising because these technologies are used to measure values that relate to lameness, in both clinical and experimental setups. However, there are many other practical uses of equine quantitative gait that can be used in clinics and training centres with the goal of improving equine health and performance. Some may require more, and others less, advanced equipment and expertise, but all current technologies are aiming to become more user-friendly for the equine practitioner. Quantitative gait analysis can identify and measure potential gait disturbances in horses with musculoskeletal or neurological conditions. Its use in aiding lameness diagnosis and monitoring is mainly based on measuring gait asymmetries, determining whether those asymmetries are over pre-set thresholds and deciding on their clinical relevance. Much work has already been done to investigate symmetry and asymmetry values of the motion of various body parts (including vertical displacement of head, withers and pelvis; joint angle parameters of the limbs and vertebral column), stride parameters and limb loading; as well as the symmetry values seen under different conditions such as different disciplines, training, type of terrain, with or without a rider and during lungeing or when moving over a straight line [1-7]. Further refinement of knowledge on gait parameters is still warranted, and research in the area will not cease as technologies evolve. Although perfectly suited for the purpose, quantitative gait analysis has been used in very few studies of neurological conditions. Gait parameters relevant to ataxia have been successfully measured in an experimental study [8]. In a kinematic study with clinically ataxic and non-ataxic horses, Olsen et al. [9] showed that motion capture equipment can objectively aid the assessment of horses with ataxia. These studies show that this type of analysis has the potential to aid diagnosis in neurological cases and monitor the impact of their treatment. Quantitative gait analysis can be applied in the evaluation of training and exercise programmes, and sport-specific biomechanics. Investigating the biomechanical horse-rider interaction in international endurance competition, two accelerometers recorded the vertical displacements of horse and rider to analyse riding techniques during different gaits in elite and advanced horse-rider dyads. Advanced duos maintained a more stable speed and elite duos increased the use of the sitting canter, while both groups used 2-point trot less often than other strategies. The study highlights the great potential of accelerometer-based wearables to characterise gait and riding strategies [1]. Using also accelerometer-based devices, Heim et al. [2] established the normal values for latero-medial and dorso-ventral range of motion (ROM) of the vertebral column in 27 Franches-Montagnes stallions and compared it with a group of mix-bred horses trotting under different exercising conditions. Variability ...

show abstract

A simple method for equine kinematic gait event detection

Abstract: This study proposes a simple kinematic method of detecting hoof-on, hoof-off and GRFz at walk, trot and canter. Further work should focus on validating the methodology in a larger number of horses and should extend the method for use on surfaces with varying levels of compliance.

Cited by 9 publications

References 14 publications

Improving gait classification in horses by using inertial measurement unit (IMU) generated data and machine learning

Improving gait classification in horses by using inertial measurement unit (IMU) generated data and machine learning

The effect of cut-off frequency when high-pass filtering equine sEMG signals during locomotion

Practical uses of quantitative gait analysis in horses

Contact Info

Product

Resources

About