Relationship Between Magnitude and Direction of Asymmetries in Facial and Limb Traits in Horses and Ponies

Lesniak, Kirsty; Williams, Jane

doi:10.1016/j.jevs.2020.103195

Cited by 2 publications

(4 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Random guessing of left and right should result in 50% correctly assigned (assuming no other misclassifications), however, our best model detected side 88.4% correctly, implying that 38.4 times out of the remaining 50 times the model improved on chance based on anatomical information alone. It is somewhat surprising that laterality detection was better than chance, however, asymmetry of limb bone dimensions is recognized in humans 24 and Equidae , 25 which our model appears to give weight to in its radiographic view classification. Horses worldwide are typically handled from their left (‘near’) side and many racehorses do fast track work in one direction only (clockwise in Hong Kong), but limb asymmetry also exists in horses not performing single‐direction training 25,26 .…”

Section: Discussionmentioning

confidence: 72%

“…It is somewhat surprising that laterality detection was better than chance, however, asymmetry of limb bone dimensions is recognized in humans 24 and Equidae , 25 which our model appears to give weight to in its radiographic view classification. Horses worldwide are typically handled from their left (‘near’) side and many racehorses do fast track work in one direction only (clockwise in Hong Kong), but limb asymmetry also exists in horses not performing single‐direction training 25,26 . Previous studies have reported significant differences between right and left limbs; for example, difference in length between right and left third metacarpal bone from Thoroughbred racehorses in Victoria and South Australia 27 ; five postmortem measurements were larger in the left femur compared to the right femur from Thoroughbred racehorses in New South Wales, Australia 28 .…”

Section: Discussionmentioning

confidence: 72%

“…Horses worldwide are typically handled from their left (‘near’) side and many racehorses do fast track work in one direction only (clockwise in Hong Kong), but limb asymmetry also exists in horses not performing single‐direction training. 25 , 26 Previous studies have reported significant differences between right and left limbs; for example, difference in length between right and left third metacarpal bone from Thoroughbred racehorses in Victoria and South Australia 27 ; five postmortem measurements were larger in the left femur compared to the right femur from Thoroughbred racehorses in New South Wales, Australia. 28 Images were supplied individually in random order and not as single‐animal sets or left–right pairs, so direct within‐individual comparisons were not available to the model to use in laterality discrimination.…”

Section: Discussionmentioning

confidence: 95%

“…While size may be a contributing factor, yet others such as length–width aspect ratio, cortical thickness or opacity, or trabecular bone texture might also contribute to the model weights. Individual horses commonly have bone dimensions that are asymmetric in the opposite direction to mean population‐level asymmetry, 25 , 26 which may help to account for our model's imperfect accuracy in laterality classification.…”

Section: Discussionmentioning

confidence: 99%

See 3 more Smart Citations

Classification of racehorse limb radiographs using deep convolutional neural networks

Silva

Mishra

Riggs

et al. 2023

Veterinary Record Open

View full text Add to dashboard Cite

Purpose To assess the capability of deep convolutional neural networks to classify anatomical location and projection from a series of 48 standard views of racehorse limbs. Materials and methods Radiographs ( N = 9504) of horse limbs from image sets made for veterinary inspections by 10 independent veterinary clinics were used to train, validate and test (116, 40 and 42 radiographs, respectively) six deep learning architectures available as part of the open source machine learning framework PyTorch. The deep learning architectures with the best top‐1 accuracy had the batch size further investigated. Results Top‐1 accuracy of six deep learning architectures ranged from 0.737 to 0.841. Top‐1 accuracy of the best deep learning architecture (ResNet‐34) ranged from 0.809 to 0.878, depending on batch size. ResNet‐34 (batch size = 8) achieved the highest top‐1 accuracy (0.878) and the majority (91.8%) of misclassification was due to laterality error. Class activation maps indicated that joint morphology, not side markers or other non‐anatomical image regions, drove the model decision. Conclusions Deep convolutional neural networks can classify equine pre‐import radiographs into the 48 standard views including moderate discrimination of laterality, independent of side marker presence.

show abstract

Section: Discussionmentioning

confidence: 72%

Section: Discussionmentioning

confidence: 72%

Section: Discussionmentioning

confidence: 95%

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

Classification of racehorse limb radiographs using deep convolutional neural networks

Silva

Mishra

Riggs

et al. 2023

Veterinary Record Open

View full text Add to dashboard Cite

show abstract

Repeatability of ultrasonographic measurement of cross‐sectional area of the multifidus muscle in the horse

Symonds,

Dart,

Perkins

et al. 2024

Equine Veterinary Education

View full text Add to dashboard Cite

SummaryBackgroundRecent studies have described techniques to acquire ultrasound measurements of the cross‐sectional area of the multifidus muscle to investigate back pain in horses and monitor the response to rehabilitation. These techniques have been extrapolated from human research but have not been clinically validated in the horse.ObjectiveThe objective of the study was to investigate whether measuring the cross‐sectional area of the multifidus muscle in horses using ultrasound is a repeatable technique.Study designCross‐sectional.MethodsUltrasound images of the articular process view and interspinous view of the multifidus muscle were acquired at T15, T18, L2 and L5 on the left and right sides. Three images were acquired at each location and each image was measured three times to give an average cross‐sectional area. Study 1 investigated the repeatability of acquiring images in a single session, and then on two separate days. Study 2 compared measurements of images acquired with a console (Philips EpiQ) and a portable (MyLab™Alpha) ultrasound machine. Study 3 investigated the repeatability of measuring the cross‐sectional area using external software, OsiriX. Intraclass correlation coefficients (ICC) were calculated for each study.ResultsRepeatability varied from good to excellent for images taken in a single session. The repeatability for measuring the images using external software was good to excellent. There was poor‐to‐moderate repeatability for images taken on two different days and poor‐to‐good repeatability when using the two different machines.Main limitationsPopulation size, study population limited to a single breed, only one operator.ConclusionsThe measurement of the multifidus muscle in horses has poor repeatability and may have limited application.

show abstract

Relationship Between Magnitude and Direction of Asymmetries in Facial and Limb Traits in Horses and Ponies

Cited by 2 publications

References 28 publications

Classification of racehorse limb radiographs using deep convolutional neural networks

Classification of racehorse limb radiographs using deep convolutional neural networks

Repeatability of ultrasonographic measurement of cross‐sectional area of the multifidus muscle in the horse

Contact Info

Product

Resources

About