Animal Activity Recognition From Sensor Data Using Ensemble Learning

Animal activity recognition is an important task to monitor the behavior of animals to know their health condition and psychological state. To provide a solution for this need, this study is aimed to build an internet of things (IoT) system that predicts the activities of animals based on sensor data obtained from embedded devices attached to animals. This chapter especially considers the problem of prediction of goat activity using three types of sensors: accelerometer, gyroscope, and magnetometer. Five possible goat activities are of interest, including stationary, grazing, walking, trotting, and running. The utility of five ensemble learning methods was investigated, including random forest, extremely randomized trees, bagging trees, gradient boosting, and extreme gradient boosting. The results showed that all these methods achieved good performance (>94%) on the datasets. Therefore, this study can be successfully used by professionals such as farmers, vets, and animal behaviorists where animal tracking may be crucial.

Sperm Quality and Histology of the Testis and Epididymis in Chemical Castrated Male Cats with Intra-testicular Injection of Sodium Chloride Solution

The high population of feral cats in Bali, Indonesia, could be a problem associated with cat welfare and the risk of zoonotic diseases. Gonadectomy or male castration is an option for population control; however, there would be more economical ways to reach this goal. Sterilization using chemicals is an alternative to the surgical method of castration as it is reasonably priced, activated in a short time, and does not affect animal activity after the chemical administration. The present study aimed to evaluate intratesticular injection of Sodium chloride as an agent of chemosterilants in male cats. In the current study, 16 healthy male cats aged nearly one year old were randomly selected. The male cats were allocated into four groups containing four cats in each. Group 1 (control) received bilateral intratesticular injection with 0.25 mL normal saline solution containing 2% Lidocaine. The cats in groups 2, 3, and 4 received bilateral intratesticular injection of 0.25 mL Sodium chloride solution containing 2% Lidocaine with a concentration of 15%, 20%, and 25%, respectively. At the end of the study (after 30 days of injection), the cats were castrated and their testes were removed for sperm quality and histopathological evaluation. The results of the present study indicated that intratesticular injection of Sodium chloride significantly decreased the motility and viability rates, and increased the abnormalities of the sperms. Histopathological studies revealed marked depletion of spermatozoa in the testes and seminiferous tubular degeneration, seminiferous tubular atrophy, and epithelial cell vacuolation. In conclusion, the injection of 20-25% solution of Sodium chloride in testes had the potential as a chemosterilant for chemical sterilization in cats.

Water sources aggregate parasites with increasing effects in more arid conditions

Shifts in landscape heterogeneity and climate can influence animal movement in ways that profoundly alter disease transmission. Water sources that are foci of animal activity have great potential to promote disease transmission, but it is unknown how this varies across a range of hosts and climatic contexts. For fecal-oral parasites, water resources can aggregate many different hosts in small areas, concentrate infectious material, and function as disease hotspots. This may be exacerbated where water is scarce and for species requiring frequent water access. Working in an East African savanna, we show via experimental and observational methods that water sources increase the density of wild and domestic herbivore feces and thus, the concentration of fecal-oral parasites in the environment, by up to two orders of magnitude. We show that this effect is amplified in drier areas and drier periods, creating dynamic and heterogeneous disease landscapes across space and time. We also show that herbivore grazing behaviors that expose them to fecal-oral parasites often increase at water sources relative to background sites, increasing potential parasite transmission at these hotspots. Critically, this effect varies by herbivore species, with strongest effects for two animals of concern for conservation and development: elephants and cattle.

Exercise-induced changes in climbing performance

Exercise is recommended to promote health and prevent a range of diseases. However, how exercise precipitates these benefits is unclear, nor do we understand why exercise responses differ so widely between individuals. We investigate how climbing ability in Drosophila melanogaster changes in response to an exercise treatment. We find extensive variation in baseline climbing ability and exercise-induced changes ranging from −13% to +20% in climbing ability. Climbing ability, and its exercise-induced change, is sex- and genotype-dependent. GWASs implicate ‘cell–cell signalling’ genes in the control of climbing ability. We also find that animal activity does not predict climbing ability and that the exercise-induced climbing ability change cannot be predicted from the activity level induced by the exercise treatment. These results provide promising new avenues for further research into the molecular pathways controlling climbing activity and illustrate the complexities involved in trying to predict individual responses to exercise.

Studying the Activity of Leaf-Litter Fauna: A Small World to Discover

Hundreds of thousands of little creatures live in soils. Some eat live plants, live animals, or both. Others, called decomposers, consume dead plants, and the waste of other living beings (their feces and their dead bodies), and transform them into food for plants. The health of soils depends largely on the presence of decomposers, and thus it is necessary to study how these creatures may be affected by climate change. To this end, we built a new type of traps to catch live soil animals, which we called cul-de-sac and basket traps. Here, we show how these traps are better for studying animal activity (how much they move in the soil) compared to the most used devices to date, pitfall traps. Comparatively, our traps capture more active animals and prevent predators from killing prey inside, which will improve the accuracy of future studies all over the world.

PSII-16 Potential of sensor technology to detect changes in behavior related to health status of ewes following exposure to mold-contaminated feed

Behavior can be a valuable indicator of an animal’s health and wellness state, emphasizing the importance of monitoring changes in behavior. Sensor technologies can identify changes in animal activity, providing the potential to detect abnormal behavior associated with health status and animal welfare. The aim of this study was to investigate sheep behavior before and after illness caused by mold-contaminated feed using tri-axial accelerometers. Twenty-five ewes were fitted with HerdDogg biometric accelerometer ear tags. Thirteen ewes also had Axivity AX3 accelerometers attached on the opposite side ear tag. Tri-axial accelerometers were attached to mature ewes for the purpose of parturition evaluation, when an unexpected ration change occurred. The entire sheep flock was exposed to mold-contaminated feed, symptoms were immediately observed, and symptomatic ewes were treated with a veterinarian directed protocol. Accelerometer data were evaluated 4 days prior to exposure (d -4 to -1); on day of ration change (d 0); and 4 days post exposure (d 1 to 4). Herddogg accelerometer data from 3 ewes displaying symptoms identified a decrease in activity (P ≤ 0.03) for 2 days after feed exposure compared to the 4 days before exposure. Three days after exposure and 2 days after treatment, no difference in activity was detected (P > 0.05) between pre- and post-activity levels. Herddogg accelerometers can identify similar behavioral changes as Axivity accelerometers. The correlation of the standard deviation of movement intensity derived from the Axivity accelerometer and the Herddogg activity metric was 0.67. It is crucial to detect subtle behavioral changes at onset of illness, to allow for prompt producer intervention before animal health is compromised. These results suggest that real-time or near-real time accelerometers, such as Herddogg, have potential to detect illness in ewes in a pen setting.

Constitutional types of horses and resources for their homeopathic correction

Selective breeding is in vogue since centuries to improve horse’ breed. It aims at production of horses with special performance and bodily characteristics. It is a complex, costly and time-consuming occupation. But the characteristics of horses depends upon market demands. Many a times a horse, which possesses planned kineto, behavior, exterior characteristics, is not in demand anymore. For many years we performed horse “homeopathic correction” studies at ‘Griffon’ Stud Farm and Horse Club on the basis of A.A. Komissarenko’s theoretical developments. We have established that not only physical characteristics of studs are inherited, but also their psychic traits, which fit into a certain constitutional type of the stud. Among such different types it is possible to identify both harmonious and disharmonious type of animals. It has been established the homeopathic remedy regimen depends on the age of an animal. In our experience, efficient homeopathic correction should be performed once in three months during the entire life-time of a horse, to maintain their bodily and psychic characters. It is advisable to differentiate the constitutional characteristics of growing horses before fence training, as it may make it easier for the horse to grow and mature. In this case young horses attain good standard, have good height and psychic and physical characteristics. At the farm we performed homeopathic correction of pregnant animals once in the eight month of pregnancy that is at three months before colt delivery. Four colts were born, which were rated as top grade on the zootechnic scale of All-Russia’s Horse-breeding Research Institute (VNIIKВ). The evaluation is based on the animal activity at different times, during the first three years of the life. In our observation, colts born after homeopathic correction in the pregnancy period present better exterior and behavioural characteristics than others in their ancestor line. The outcome of our research is the conclusion that in horse breeding it makes sense to have early homeopathic correction. It improves animal’s performance, the rapport between horses and men and ensures horse’s career success. Achievements in production of impeccable growing horses were highly appreciated by Trakehner Union in Russia, and the Griffon Stud Farm became a full-fledged member of the Union in 2015. Utilization of homeopathy resources allows advance of zootechnic industry, as well as prompt response to market demand; the earlier such work begins, the more efficient and cost-effective it becomes.

Exercise-induced changes in climbing performance

Exercise is recommended to promote health and prevent a range of diseases. However, how exercise precipitates these benefits is unclear, nor do we understand why exercise responses differ so widely between individuals. We investigate how climbing ability in Drosophila melanogaster changes in response to an exercise treatment. We find extensive variation in baseline climbing ability, and exercise-induced changes ranging from -13% to +20% in climbing ability. Climbing ability, and its exercise-induced change, are sex- and genotype-dependent. GWASs implicate "cell-cell signaling" genes in the control of climbing ability. We also find that animal activity does not predict climbing ability and that the exercise-induced climbing ability change cannot be predicted from the activity level induced by the exercise treatment. These results provide promising new avenues for further research into the molecular pathways controlling climbing activity and illustrate the complexities involved in trying to predict individual responses to exercise.

Application of YOLOv4 for Detection and Motion Monitoring of Red Foxes

Animal activity is an indicator for its welfare and manual observation is time and cost intensive. To this end, automatic detection and monitoring of live captive animals is of major importance for assessing animal activity, and, thereby, allowing for early recognition of changes indicative for diseases and animal welfare issues. We demonstrate that machine learning methods can provide a gap-less monitoring of red foxes in an experimental lab-setting, including a classification into activity patterns. Therefore, bounding boxes are used to measure fox movements, and, thus, the activity level of the animals. We use computer vision, being a non-invasive method for the automatic monitoring of foxes. More specifically, we train the existing algorithm ‘you only look once’ version 4 (YOLOv4) to detect foxes, and the trained classifier is applied to video data of an experiment involving foxes. As we show, computer evaluation outperforms other evaluation methods. Application of automatic detection of foxes can be used for detecting different movement patterns. These, in turn, can be used for animal behavioral analysis and, thus, animal welfare monitoring. Once established for a specific animal species, such systems could be used for animal monitoring in real-time under experimental conditions, or other areas of animal husbandry.