| 1. |
- Haubro Andersen, Pia, et al.
(författare)
-
Towards Machine Recognition of Facial Expressions of Pain in Horses
- 2021
-
Ingår i: Animals. - : MDPI. - 2076-2615. ; 11:6
-
Forskningsöversikt (refereegranskat)abstract
- Simple Summary Facial activity can convey valid information about the experience of pain in a horse. However, scoring of pain in horses based on facial activity is still in its infancy and accurate scoring can only be performed by trained assessors. Pain in humans can now be recognized reliably from video footage of faces, using computer vision and machine learning. We examine the hurdles in applying these technologies to horses and suggest two general approaches to automatic horse pain recognition. The first approach involves automatically detecting objectively defined facial expression aspects that do not involve any human judgment of what the expression "means". Automated classification of pain expressions can then be done according to a rule-based system since the facial expression aspects are defined with this information in mind. The other involves training very flexible machine learning methods with raw videos of horses with known true pain status. The upside of this approach is that the system has access to all the information in the video without engineered intermediate methods that have filtered out most of the variation. However, a large challenge is that large datasets with reliable pain annotation are required. We have obtained promising results from both approaches. Automated recognition of human facial expressions of pain and emotions is to a certain degree a solved problem, using approaches based on computer vision and machine learning. However, the application of such methods to horses has proven difficult. Major barriers are the lack of sufficiently large, annotated databases for horses and difficulties in obtaining correct classifications of pain because horses are non-verbal. This review describes our work to overcome these barriers, using two different approaches. One involves the use of a manual, but relatively objective, classification system for facial activity (Facial Action Coding System), where data are analyzed for pain expressions after coding using machine learning principles. We have devised tools that can aid manual labeling by identifying the faces and facial keypoints of horses. This approach provides promising results in the automated recognition of facial action units from images. The second approach, recurrent neural network end-to-end learning, requires less extraction of features and representations from the video but instead depends on large volumes of video data with ground truth. Our preliminary results suggest clearly that dynamics are important for pain recognition and show that combinations of recurrent neural networks can classify experimental pain in a small number of horses better than human raters.
|
|
| 2. |
- Rashid, Maheen, et al.
(författare)
-
Equine Pain Behavior Classification via Self-Supervised Disentangled Pose Representation
- 2022
-
Ingår i: 2022 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV 2022). - : Institute of Electrical and Electronics Engineers (IEEE). ; , s. 152-162
-
Konferensbidrag (refereegranskat)abstract
- Timely detection of horse pain is important for equine welfare. Horses express pain through their facial and body behavior, but may hide signs of pain from unfamiliar human observers. In addition, collecting visual data with detailed annotation of horse behavior and pain state is both cumbersome and not scalable. Consequently, a pragmatic equine pain classification system would use video of the un-observed horse and weak labels. This paper proposes such a method for equine pain classification by using multi-view surveillance video footage of unobserved horses with induced orthopaedic pain, with temporally sparse video level pain labels. To ensure that pain is learned from horse body language alone, we first train a self-supervised generative model to disentangle horse pose from its appearance and background before using the disentangled horse pose latent representation for pain classification. To make best use of the pain labels, we develop a novel loss that formulates pain classification as a multi-instance learning problem. Our method achieves pain classification accuracy better than human expert performance with 60% accuracy. The learned latent horse pose representation is shown to be viewpoint covariant, and disentangled from horse appearance. Qualitative analysis of pain classified segments shows correspondence between the pain symptoms identified by our model, and equine pain scales used in veterinary practice.
|
|
| 3. |
- Hernlund, Elin, et al.
(författare)
-
Comparing subjective and objective evaluation of show jumping competition and warm-up arena surfaces
- 2017
-
Ingår i: Veterinary Journal. - : Elsevier BV. - 1090-0233 .- 1532-2971. ; 227, s. 49-57
-
Tidskriftsartikel (refereegranskat)abstract
- The development of safety and quality standards for equestrian surfaces needs to be based on objective, repeatable measurements which allow comparisons between surfaces. These measurements should incorporate the assessment of surface performance by riders. This study provides data from objective and subjective assessment of functional properties of high-level show jumping competition and warm-up arenas. Twenty-five arenas in nine international show jumping events were evaluated by mechanical in situ testing with a surface tester, rider assessments using visual analogue scales (198 riders provided 749 arena evaluations), descriptions of arena constructions and by laboratory tests of surface material. Mixed models were used to present subjective evaluation of rider perception of the functional properties for each arena while controlling for rider and event. The association between objective and subjective assessments were also explored creating mixed models, controlling for rider and event.Mechanical measurements of impact firmness, and to a lesser extent cushioning and grip, had a significant positive association with the riders' perception. Responsiveness as assessed by the Orono biomechanical surface tester (OBST) was negatively associated with the riders' perceptions, which suggests riders and the OBST had different concepts of this functional property and that further developments of the OBST might be necessary. Objectively measured uniformity showed no useful association with riders' perception. Even though arena assessments were made by top level riders, a substantial inter-rider variation was demonstrated. (C) 2017 The Authors. Published by Elsevier Ltd.
|
|
| 4. |
- Egenvall, Agneta, et al.
(författare)
-
Days-lost to training and competition in relation to workload in 263 elite show-jumping horses in four European countries
- 2013
-
Ingår i: Preventive Veterinary Medicine. - : Elsevier BV. - 0167-5877 .- 1873-1716. ; 112, s. 387-400
-
Tidskriftsartikel (refereegranskat)abstract
- Orthopaedic, or other, injuries in sports medicine can be quantified using the ‘days-lost to training’ concept. Both the training regimen and the surface used in training and racing can affect the health of racehorses. Our aim was to associate ‘days-lost to training’ in elite-level show-jumpers to horse characteristics, training and management strategies, and the time spent working on various training and competition surfaces. We designed a longitudinal study of professional riders in four European countries. Data were recorded using training diaries. Reasons for days-lost were classified into non-acute and acute orthopaedic, medical, hoof-related, and undefined. We produced descriptive statistics of training durations, relative to type of training, surfaces used, and days-lost. We created zero-inflated negative-binomial random-effects models using the overall days-lost as outcome. In the whole dataset, duration variables related to training surfaces were analysed as independent. The Swedish data only were also used to test whether duration variables were related to competition surfaces. Thirty-one riders with 263 horses provided data on 39,028 days at risk. Of these, 2357 (6.0%) were days-lost (55% and 22% of these were due to non-acute and acute orthopaedic injuries, respectively) in 126 horses. In the all-country model, controlling for season, a significant variable was country. Switzerland and the UK had lower incidence-rate ratios (IR) compared to Sweden (IRs 0.2 and 0.03, respectively). Horses with previous orthopaedic problems had almost a doubled IR (1.8) of days-lost due to orthopaedic injury, compared to baseline. If the horse had jumping training more than 1 minute per day at risk the IRs were 6.9-7 (compared to less than this amount of time); this was, however, likely an effect of a small baseline. Variation in training was a protective factor with a dose-response relationship; the category with the highest variation had an IR of 0.1. In the Swedish model, controlling for season, there was an association of year (IR 2.8 year 2010). Further, if the horse rested >17-25% of the days at risk, or >33% of the DAR2, had IRs 3.5 and 3,0, compared to less time. Horses ≥6 years had IRs of 1.8-2.0, compared to younger horses. Limited training use of sand surface was a risk-factor (IR 2.2; >4≤12 min/day at risk), compared to not training on sand. Training/competing on sand-wood was a protective factor (IRs 0.4-0.5) compared to not using this surface.
|
|
| 5. |
- Lawin, Felix Jaremo, et al.
(författare)
-
Is Markerless More or Less? : Comparing a Smartphone Computer Vision Method for Equine Lameness Assessment to Multi-Camera Motion Capture
- 2023
-
Ingår i: Animals. - : MDPI AG. - 2076-2615. ; 13:3
-
Tidskriftsartikel (refereegranskat)abstract
- Lameness, an alteration of the gait due to pain or dysfunction of the locomotor system, is the most common disease symptom in horses. Yet, it is difficult for veterinarians to correctly assess by visual inspection. Objective tools that can aid clinical decision making and provide early disease detection through sensitive lameness measurements are needed. In this study, we describe how an AI-powered measurement tool on a smartphone can detect lameness in horses without the need to mount equipment on the horse. We compare it to a state-of-the-art multi-camera motion capture system by simultaneous, synchronised recordings from both systems. The mean difference between the systems' output of lameness metrics was below 2.2 mm. Therefore, we conclude that the smartphone measurement tool can detect lameness at relevant levels with easy-of-use for the veterinarian. 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.
|
|
| 6. |
- Persson-Sjödin, Emma, et al.
(författare)
-
Effect of meloxicam treatment on movement asymmetry in riding horses in training
- 2019
-
Ingår i: PLOS ONE. - : Public Library of Science (PLoS). - 1932-6203. ; 14:8
-
Tidskriftsartikel (refereegranskat)abstract
- Quantitative gait analysis has revealed that a large proportion of horses in training, perceived as free from lameness by their owners, show movement asymmetries of equal magnitude to horses with mild clinical lameness. Whether these movement asymmetries are related to orthopaedic pain and/or pathology has yet to be further investigated. Therefore, the objective of this study was to determine whether movement asymmetries in riding horses in training are affected by anti-inflammatory treatment with meloxicam. In a crossover design, horses were treated with meloxicam or placebo for four days respectively, with a 14-16 day washout period between treatments. Objective movement analysis utilising body mounted accelerometers was performed on a hard and a soft surface before and on day four of each treatment. A trial mean was calculated for the differences between the two vertical displacement minima and maxima of head (HDmin, HDmax) and pelvis (PDmin, PDmax) per stride. Horses (n = 66) with trial mean asymmetries greater than 6 mm for HDmin or HDmax, or more than 3 mm for PDmin or PDmax, at baseline were included. The difference before and after each treatment in the measured movement asymmetry was assessed with linear mixed models. Treatment with meloxicam did not significantly affect the movement asymmetry in any of the models applied (all p>0.30). These results raise new questions: are the movement asymmetries in riding horses in training simply expressions of biological variation or are they related to pain/dysfunction that is non-responsive to meloxicam treatment?
|
|
| 7. |
- Rhodin, Marie, et al.
(författare)
-
Timing of Vertical Head, Withers and Pelvis Movements Relative to the Footfalls in Different Equine Gaits and Breeds
- 2022
-
Ingår i: Animals. - : MDPI AG. - 2076-2615. ; 12
-
Tidskriftsartikel (refereegranskat)abstract
- Simple Summary Movement symmetry of the head and pelvis are used to measure lameness in horses in trot. Although head, pelvis and limb movements have been described, less is known about the temporal relationships between them. This information is needed to understand how the movements change with lameness. This is particularly relevant in gaited horses, such as the Icelandic horse that perform gaits such as tolt and pace, which are challenging to evaluate. This study used inertial measurement units to investigate head, withers and pelvis motion relative to limb movements in Icelandic, Warmblood and Iberian horses. Limb movements, together with vertical movements and lowest/highest positions of the head, withers and pelvis were calculated, and the relative timing of the events was compared across breeds. Additionally, data for tolt and pace were collected and evaluated in ridden Icelandic horses. For all gaits except walk and pace, the lowest/highest positions of the head/withers/pelvis were closely temporally related to midstance and hoof-off, respectively. Pelvic and withers total range of motion differed between all breeds. The Icelandic horses showed shorter stride duration and smaller movements of the upper body than the other breeds at trot, which may explain why lameness evaluation in this breed is challenging. Knowledge of vertical motion patterns of the axial body segments is a prerequisite for the development of algorithms used in automated detection of lameness. To date, the focus has been on the trot. This study investigates the temporal synchronization between vertical motion of the axial body segments with limb kinematic events in walk and trot across three popular types of sport horses (19 Warmbloods, 23 Iberians, 26 Icelandics) that are known to have different stride kinematics, and it presents novel data describing vertical motion of the axial body segments in tolting and pacing Icelandic horses. Inertial measurement unit sensors recorded limb kinematics, vertical motion of the axial body at all symmetrical gaits that the horse could perform (walk, trot, tolt, pace). Limb kinematics, vertical range of motion and lowest/highest positions of the head, withers and pelvis were calculated. For all gaits except walk and pace, lowest/highest positions of the pelvis and withers were found to be closely related temporally to midstance and start of suspension of the hind/fore quarter, respectively. There were differences in pelvic/withers range of motion between all breeds where the Icelandic horses showed the smallest motion, which may explain why lameness evaluation in this breed is challenging.
|
|
| 8. |
- Leclercq, Anna, et al.
(författare)
-
Evaluation of feedback methods for improved detection of hindlimb lameness in horses among riding instructors and trainers
- 2022
-
Ingår i: Frontiers in Veterinary Science. - : Frontiers Media SA. - 2297-1769. ; 9
-
Tidskriftsartikel (refereegranskat)abstract
- Lameness, a wellknown issue in sport horses, impedes performance and impairs welfare. Early detection of lameness is essential for horses to receive needed treatment, but detection of hindlimb lameness is challenging. Riding instructors and trainers observe horses in motion in their daily work and could contribute to more efficient lameness detection. In this cross-sectional and prospective study, we evaluated the ability of riding instructors and trainers to assess hindlimb lameness. We also evaluated different feedback methods for improved lameness detection. For the cross-sectional part, n = 64 riding instructors and trainers of varying level and n = 23 high-level trainers were shown 13 videos of trotting horses, lameness degree: 0-3.5 (test 1) and tasked with classifying the horses as sound, left hindlimb lame, or right hindlimb lame. For the prospective part, the riding instructors and trainers of varying levels were randomly allocated to three different groups (a, b, c) and given 14 days of feedback-based, computer-aided training in identifying hindlimb lameness, where they assessed 13 videos (of which three were repeated from test 1) of horses trotting in a straight line. Participants in groups a-c received different feedback after each video (group a: correct answer and re-viewing of video at full and 65% speed; group b: correct answer, re-viewing of video at full and 65% speed, narrator providing explanations; group c: correct answer and re-viewing of video at full speed). After computer-aided training, the participants were again subjected to the video test (test 2). Participants also provided background information regarding level of training etcetera. Effects of participants' background on results were analyzed using analysis of variance, and effects of the different feedback methods were analyzed using generalized estimation equations. On test 1, 44% (group a), 48% (b), 46% (c), and 47% (high-level trainers) of horses were correctly classified. Group a participants significantly improved their test score, both with (p < 0.0001) and without (p = 0.0086) inclusion of repeated videos. For group c, significant improvement was only seen with inclusion of repeated videos (p = 0.041). For group b, no significant improvement was seen (p = 0.51). Although test 2 scores were low, computer-aided training may be useful for improving hindlimb lameness detection.
|
|
| 9. |
- Li, Ci, et al.
(författare)
-
The Poses for Equine Research Dataset (PFERD)
- 2024
-
Ingår i: Scientific Data. - : Springer Nature. - 2052-4463. ; 11:1
-
Tidskriftsartikel (refereegranskat)abstract
- Studies of quadruped animal motion help us to identify diseases, understand behavior and unravel the mechanics behind gaits in animals. The horse is likely the best-studied animal in this aspect, but data capture is challenging and time-consuming. Computer vision techniques improve animal motion extraction, but the development relies on reference datasets, which are scarce, not open-access and often provide data from only a few anatomical landmarks. Addressing this data gap, we introduce PFERD, a video and 3D marker motion dataset from horses using a full-body set-up of densely placed over 100 skin-attached markers and synchronized videos from ten camera angles. Five horses of diverse conformations provide data for various motions from basic poses (eg. walking, trotting) to advanced motions (eg. rearing, kicking). We further express the 3D motions with current techniques and a 3D parameterized model, the hSMAL model, establishing a baseline for 3D horse markerless motion capture. PFERD enables advanced biomechanical studies and provides a resource of ground truth data for the methodological development of markerless motion capture.
|
|
| 10. |
- Persson-Sjodin, Emma, et al.
(författare)
-
Withers vertical movement symmetry is useful for locating the primary lame limb in naturally occurring lameness
- 2024
-
Ingår i: Equine Veterinary Journal. - : WILEY. - 0425-1644 .- 2042-3306. ; 56:1, s. 76-88
-
Tidskriftsartikel (refereegranskat)abstract
- BackgroundDuring orthopaedic assessment of lame horses, a head nod is commonly present in both primary forelimb and hindlimb lame horses. Additional motion metrics that could assist clinicians in correctly differentiating between these two scenarios would be of great clinical value. ObjectivesThe primary objective of this study was to examine whether withers movement asymmetry can be used in a clinical setting to distinguish primary forelimb lameness from compensatory head movement asymmetry due to primary hindlimb lameness. Study designRetrospective, multicentre study. MethodsMovement asymmetry of head, withers and pelvis was measured using multi-camera optical motion capture, as part of routine lameness investigations at four European equine hospitals. Vertical movement asymmetry parameters from 317 horses trotting in a straight line were compared before and after successful diagnostic analgesia of a single limb. Descriptive statistics, t-tests and linear models were used to analyse the data. ResultsIn forelimb lame horses, 80%-81% showed head and withers asymmetry both indicating lameness in the same forelimb. In hindlimb lame horses, 69%-72% showed head asymmetry ipsilateral to the lame hindlimb and withers asymmetry diagonal to the lame hindlimb, thus, head and withers asymmetry indicated lameness in different forelimbs. A large (>15 mm) compensatory head nod was seen in 28%-31% of the hindlimb lame horses. In 89%-92% of these, head and withers asymmetry indicated lameness in different forelimbs. Withers asymmetry decreased linearly with reduced head or pelvic asymmetry for both forelimb and hindlimb lame horses. Main limitationsCompensatory strategies were evaluated on group level to identify common patterns, potentially ignoring uncommon individual strategies. ConclusionsWithers vertical movement asymmetry metrics can be useful in helping to locate the primary lame limb during quantitative lameness assessment. Head and withers movement asymmetry parameters generally indicate the same forelimb in forelimb lame horses, but different forelimbs in hindlimb lame horses.
|
|
