Integrating Audio Signal Processing and Deep Learning Algorithms for Gait Pattern Classification in Brazilian Gaited Horses

This study focused on assessing the usefulness of using audio signal processing in the gaited horse industry. A total of 196 short-time audio files (4 s) were collected from video recordings of Brazilian gaited horses. These files were converted into waveform signals (196 samples by 80,000 columns) and divided into training (N = 164) and validation (N = 32) datasets. Twelve single-valued audio features were initially extracted to summarize the training data according to the gait patterns (Marcha Batida—MB and Marcha Picada—MP). After preliminary analyses, high-dimensional arrays of the Mel Frequency Cepstral Coefficients (MFCC), Onset Strength (OS), and Tempogram (TEMP) were extracted and used as input information in the classification algorithms. A principal component analysis (PCA) was performed using the 12 single-valued features set and each audio-feature dataset—AFD (MFCC, OS, and TEMP) for prior data visualization. Machine learning (random forest, RF; support vector machine, SVM) and deep learning (multilayer perceptron neural networks, MLP; convolution neural networks, CNN) algorithms were used to classify the gait types. A five-fold cross-validation scheme with 10 repetitions was employed for assessing the models' predictive performance. The classification performance across models and AFD was also validated with independent observations. The models and AFD were compared based on the classification accuracy (ACC), specificity (SPEC), sensitivity (SEN), and area under the curve (AUC). In the logistic regression analysis, five out of the 12 audio features extracted were significant (p < 0.05) between the gait types. ACC averages ranged from 0.806 to 0.932 for MFCC, from 0.758 to 0.948 for OS and, from 0.936 to 0.968 for TEMP. Overall, the TEMP dataset provided the best classification accuracies for all models. The most suitable method for audio-based horse gait pattern classification was CNN. Both cross and independent validation schemes confirmed that high values of ACC, SPEC, SEN, and AUC are expected for yet-to-be-observed labels, except for MFCC-based models, in which clear overfitting was observed. Using audio-generated data for describing gait phenotypes in Brazilian horses is a promising approach, as the two gait patterns were correctly distinguished. The highest classification performance was achieved by combining CNN and the rhythmic-descriptive AFD.

Thermographic quantitative analysis by quadrants of australian saddles used in basic horseback riding courses

The saddle is a particularly important piece of equipment when it comes to training horses. Any problem with this, or in its adjustment, can cause discomfort, pain and injury to the horses. The aim of this study was to analyze Australian saddles thermographically, in a quantitative way by quadrants, performing a primary detection of how pressure distribution occurs in Australian saddles commonly used in Mangalarga Marchador (MM) gaited horses during a basic riding course, in order to signal whether there is a need for a more complete assessment of saddle befitting this type of activity. For this purpose, ten similar Australian saddles used by ten MM horses were thermographically evaluated during a basic riding course. Thermographic images were obtained from the ventral face of the saddle before and after one hour of exercise. The images were divided into 9 quadrants and compared quantitatively before and after the ride. Quantitative analysis by quadrants including 3 central quadrants for assessing the gullet region was effective, providing the necessary detail for data analysis. Inappropriate contact of the gullet region with the animal’s back was demonstrated in 100% of the saddles, with the central quadrant of the gullet being the most affected, indicating the inadequacy of this type of saddle for MM in this activity. Therefore, there is a need for a more complete assessment of saddle befitting an Australian saddle type to this equestrian activity, aiming at improvements in animal welfare.

Use of oil-rich diet for gaited horses during physical training

In Brazil, gaited horses are selected based on field tests, during which they move at speeds of 3 to 4 m/s for 30 to 60 min. To cover their nutrient requirements, feed manufacturers have developed oil-rich dietary supplements and concentrates. The aim of this research was to evaluate the effects of increasing the density of dietary fats in the feed of 16 Campolina horses undergoing intense gaited training. The training schedule consisted of training 4 × a week: 3 days riding for 60 min (10' warm-up, 40' doing marcha gait, and 10' cool down, and 1 day walking for 90'). The horses were divided in two groups: control and supplemented. The dietary supplementation was isocaloric, with the control group receiving an ordinary concentrate (3.5% fat) and the supplemented group fed 1.0 kg of the supplement (18% fat) plus the ordinary concentrate. Both groups had free access to fresh grass, salt and water. Blood samples were collected prior to and after 4 and 8 weeks of supplementation to calculate the complete blood count, glucose, total protein, triglycerides, total cholesterol, HDL, LDL and non-esterified fatty acids (NEFA). The results were evaluated by ANOVA and Tukey's test (P < 0.05). The supplemented group showed elevated levels of NEFA, red blood cells and haematocrit (P < 0.05), while the control group showed high triglyceride concentrations; both groups showed reduced plasma protein concentrations (P < 0.05). The other indices did not change (P > 0.05). The increase in dietary fat intake increased the blood lipid biomarkers and antioxidant capacity of gaited horses during intense training, possibly contributing to improve their metabolic performance.