Accurate objective, automated limb lameness detection and pose estimation play an important role for animal well-being and precision livestock farming. We present a wearable sensor-based limb lameness detection and pose estimation for horse walk and trot locomotion. The gait event and lameness detection are first built on a recurrent neural network (RNN) with long short-term memory (LSTM) cells. Its outcomes are used in the limb pose estimation. A learned low-dimensional motion manifold is parameterized by a phase variable with a Gaussian process dynamic model. We compare the RNN-LSTM-based lameness detection method with a feature-based multi-layer classifier (MLC) and a multi-class classifier (MCC) that are built on support vector machine/K-nearest-neighbors and deep convolutional neural network methods, respectively. Experimental results show that using only accelerometer measurements, the RNN-LSTM-based approach achieves 95% lameness detection accuracy and also outperforms the feature-based MLC or MCC in terms of several assessment criteria. The pose estimation scheme can predict the 24 limb joint angles in the sagittal plane with average errors less than 5 and 10 degs under normal and induced lameness conditions, respectively. The presented work demonstrate the successful use of machine learning techniques for high performance lameness detection and pose estimation in equine science. Note to Practitioners —Automation technologies are increasingly used for precision agriculture but few have focused on monitoring individual animals in open field for precision livestock farming. Limb lameness detection and pose estimation in open field is labor-intensive, unsafe for farmers, and inefficient. The presented machine learning-enabled, wearable inertial sensor-based design provides an effective and efficient approach for horse limb lameness detection and pose estimation applications. We present an RNN-LSTM for lameness detection and an integrated manifold learning model is used to predict the horse limb joint angles in walk and trot gaits under normal and induced lameness conditions. We also present a systematic analysis and experiments to demonstrate the impacts of the wearable sensor locations and signal information on lameness detection and pose estimation performance. Several other machine learning-based lameness detection methods are also presented and compared. The extensive multi-horse testing results are presented to demonstrate the superior accuracy and higher performance than other types of machine learning methods. One attractive feature of the proposed design lies in its high performance and fast computational capability for potential real-time applications in open field.

Little published information exists on the horses in equine-assisted services (EAS), particularly their selection, longevity, and retirement. The purpose of this study was to characterize horses and procedures used in EAS. A pilot survey was developed using focus group discussions and distributed to Professional Association of Therapeutic Horsemanship, International (PATH Intl) centers in Florida (n = 45, Part I) before further modification and distribution to members of PATH Intl., American Hippotherapy Association (AHA), eagala, and Certified Horsemanship Association (CHA) (n = 26,000, Part II). Response rates were 36% (Part I) and 0.7% (Part II). Centers report a median of 10 (Part I) or 9 (Part II) horses and ponies. Selection procedures included initial screening (Part I = 100%, Part II = 96%), pre-purchase or pre-donation exam (I = 64%, II = 60%), acclimation period (I = 100%, II = 84%), trial period (I = 91%, II = 90%), and other (II = 11%). Horses remained active in programs for less than a year to over 20 years with the greatest number working 7–10 (Part I) or 1–6 (Part II) yr. In Part I of the study, behavior (44%) was the leading cause of retirement followed by unsoundness (33%). In Part II, unsoundness was the highest ranked response followed by behavior. Behavior, soundness, and health emerged as key factors in horse selection and retirement. Future work should focus on investigating these issues at an individual horse level.

Horses' muscular tension during acute stress remains unexplored. Our aim was to assess muscular, behavioral, cortisol, and hematocrit responses to social isolation (ISO), novel object exposure (NOV), and sham clipping (CLIP). Altered stress responses were expected. Eight mature Standardbred horses (four mares and four geldings) were exposed to acute stressors and a control period (CON) in a balanced. replicated 4x4 Latin Square experimental design with 3 min treatment periods and 1.0 min washout periods. Surface electromyography collected from the musseter, brachiocephalus, cervical trapezius, and longissimus dorsi was processed to derive average rectified value (ARV) and median frequency (MF) during the initial, middle, and final 30 s of treatments. ARV and MF data were log transformed then analyzed using a mixed model, repeated measures ANOVA along with plasma cortisol and hematocrit. Behavior data were analyzed using a negative binomial distribution mixed model ANOVA. CLIP resulted in greater (p < 0.05) log ARV in the masseter (1.5 + 1.5%, mean + SD) and brachiocepahlas (2.2 + 2.0%) than CON (-1.2 + 1.4%, 0.1 + 1.5%). ISO resulted in greater (p < 0.05) log ARV in the masseter (0.2 + 1.3%) and cervical trapezius (0.6 + 1.3%) than CON (-1.2 + 1.4%, -1.0 + 1.7%). ISO increased (p < 0.05) the total number of stress-related behaviors and hematocrit. No changes in cortisol were observed. We suggest that muscular tension can be used as an indicator of acute stress in horses. Incorporating muscle activity into an array of measurements may provide a more nuanced understanding of stress responses.