Early detection of health and welfare compromises through automated detection of behavioural changes in pigs

Early detection of bovine respiratory disease in pre-weaned dairy calves using sensor based feeding, movement, and social behavioural data

Previous research shows that feeding and activity behaviours in combination with machine learning algorithms has the potential to predict the onset of bovine respiratory disease (BRD). This study used 229 novel and previously researched feeding, movement, and social behavioural features with machine learning classification algorithms to predict BRD events in pre-weaned calves. Data for 172 group housed calves were collected using automatic milk feeding machines and ultrawideband location sensors. Health assessments were carried out twice weekly using a modified Wisconsin scoring system and calves were classified as sick if they had a Wisconsin score of five or above and/or a rectal temperature of 39.5 °C or higher. A gradient boosting machine classification algorithm produced moderate to high performance: accuracy (0.773), precision (0.776), sensitivity (0.625), specificity (0.872), and F1-score (0.689). The most important 30 features were 40% feeding, 50% movement, and 10% social behavioural features. Movement behaviours, specifically the distance walked per day, were most important for model prediction, whereas feeding and social features aided in the model's prediction minimally. These results highlighting the predictive potential in this area but the need for further improvement before behavioural changes can be used to reliably predict the onset of BRD in pre-weaned calves.

An Integrated Gather-and-Distribute Mechanism and Attention-Enhanced Deformable Convolution Model for Pig Behavior Recognition

The behavior of pigs is intricately tied to their health status, highlighting the critical importance of accurately recognizing pig behavior, particularly abnormal behavior, for effective health monitoring and management. This study addresses the challenge of accommodating frequent non-rigid deformations in pig behavior using deformable convolutional networks (DCN) to extract more comprehensive features by incorporating offsets during training. To overcome the inherent limitations of traditional DCN offset weight calculations, the study introduces the multi-path coordinate attention (MPCA) mechanism to enhance the optimization of the DCN offset weight calculation within the designed DCN-MPCA module, further integrated into the cross-scale cross-feature (C2f) module of the backbone network. This optimized C2f-DM module significantly enhances feature extraction capabilities. Additionally, a gather-and-distribute (GD) mechanism is employed in the neck to improve non-adjacent layer feature fusion in the YOLOv8 network. Consequently, the novel DM-GD-YOLO model proposed in this study is evaluated on a self-built dataset comprising 11,999 images obtained from an online monitoring platform focusing on pigs aged between 70 and 150 days. The results show that DM-GD-YOLO can simultaneously recognize four common behaviors and three abnormal behaviors, achieving a precision of 88.2%, recall of 92.2%, and mean average precision (mAP) of 95.3% with 6.0MB Parameters and 10.0G FLOPs. Overall, the model outperforms popular models such as Faster R-CNN, EfficientDet, YOLOv7, and YOLOv8 in monitoring pens with about 30 pigs, providing technical support for the intelligent management and welfare-focused breeding of pigs while advancing the transformation and modernization of the pig industry.

Pattern Mining-Based Pig Behavior Analysis for Health and Welfare Monitoring

The increasing popularity of pigs has prompted farmers to increase pig production to meet the growing demand. However, while the number of pigs is increasing, that of farm workers has been declining, making it challenging to perform various farm tasks, the most important among them being managing the pigs' health and welfare. This study proposes a pattern mining-based pig behavior analysis system to provide visualized information and behavioral patterns, assisting farmers in effectively monitoring and assessing pigs' health and welfare. The system consists of four modules: (1) data acquisition module for collecting pigs video; (2) detection and tracking module for localizing and uniquely identifying pigs, using tracking information to crop pig images; (3) pig behavior recognition module for recognizing pig behaviors from sequences of cropped images; and (4) pig behavior analysis module for providing visualized information and behavioral patterns to effectively help farmers understand and manage pigs. In the second module, we utilize ByteTrack, which comprises YOLOx as the detector and the BYTE algorithm as the tracker, while MnasNet and LSTM serve as appearance features and temporal information extractors in the third module. The experimental results show that the system achieved a multi-object tracking accuracy of 0.971 for tracking and an F1 score of 0.931 for behavior recognition, while also highlighting the effectiveness of visualization and pattern mining in helping farmers comprehend and manage pigs' health and welfare.

Detecting tail biters by monitoring pig screams in weaning pigs

Early identification of tail biting and intervention are necessary to reduce tail lesions and their impact on animal health and welfare. Removal of biters has become an effective intervention strategy, but finding them can be difficult and time-consuming. The aim of this study was to investigate whether tail biting and, in particular, individual biters could be identified by detecting pig screams in audio recordings. The study included 288 undocked weaner pigs housed in six pens in two batches. Once a tail biter (n = 7) was identified by visual inspection in the stable and removed by the farm staff, the previous days of video and audio recordings were analyzed for pig screams (sudden increase in loudness with frequencies above 1 kHz) and tail biting events until no biting before the removal was observed anymore. In total, 2893 screams were detected in four pens where tail biting occurred. Of these screams, 52.9% were caused by tail biting in the observed pen, 25.6% originated from other pens, 8.8% were not assignable, and 12.7% occurred due to other reasons. In case of a tail biting event, screams were assigned individually to biter and victim pigs. Based on the audio analysis, biters were identified between one and nine days prior to their removal from the pen after visual inspection. Screams were detected earlier than the increase in hanging tails and could therefore be favored as an early warning indicator. Analyzing animal vocalization has potential for monitoring and early detection of tail biting events. In combination with individual marks and automatic analysis algorithms, biters could be identified and tail biting efficiently reduced. In this way, biters can be removed earlier to increase animal health and welfare.

BovineTalk: machine learning for vocalization analysis of dairy cattle under the negative affective state of isolation

There is a critical need to develop and validate non-invasive animal-based indicators of affective states in livestock species, in order to integrate them into on-farm assessment protocols, potentially via the use of precision livestock farming (PLF) tools. One such promising approach is the use of vocal indicators. The acoustic structure of vocalizations and their functions were extensively studied in important livestock species, such as pigs, horses, poultry, and goats, yet cattle remain understudied in this context to date. Cows were shown to produce two types of vocalizations: low-frequency calls (LF), produced with the mouth closed, or partially closed, for close distance contacts, and open mouth emitted high-frequency calls (HF), produced for long-distance communication, with the latter considered to be largely associated with negative affective states. Moreover, cattle vocalizations were shown to contain information on individuality across a wide range of contexts, both negative and positive. Nowadays, dairy cows are facing a series of negative challenges and stressors in a typical production cycle, making vocalizations during negative affective states of special interest for research. One contribution of this study is providing the largest to date pre-processed (clean from noises) dataset of lactating adult multiparous dairy cows during negative affective states induced by visual isolation challenges. Here, we present two computational frameworks-deep learning based and explainable machine learning based, to classify high and low-frequency cattle calls and individual cow voice recognition. Our models in these two frameworks reached 87.2 and 89.4% accuracy for LF and HF classification, with 68.9 and 72.5% accuracy rates for the cow individual identification, respectively.

Heterogeneous fusion of biometric and deep physiological features for accurate porcine cough recognition

Accurate identification of porcine cough plays a vital role in comprehensive respiratory health monitoring and diagnosis of pigs. It serves as a fundamental prerequisite for stress-free animal health management, reducing pig mortality rates, and improving the economic efficiency of the farming industry. Creating a representative multi-source signal signature for porcine cough is a crucial step toward automating its identification. To this end, a feature fusion method that combines the biological features extracted from the acoustic source segment with the deep physiological features derived from thermal source images is proposed in the paper. First, acoustic features from various domains are extracted from the sound source signals. To determine the most effective combination of sound source features, an SVM-based recursive feature elimination cross-validation algorithm (SVM-RFECV) is employed. Second, a shallow convolutional neural network (named ThermographicNet) is constructed to extract deep physiological features from the thermal source images. Finally, the two heterogeneous features are integrated at an early stage and input into a support vector machine (SVM) for porcine cough recognition. Through rigorous experimentation, the performance of the proposed fusion approach is evaluated, achieving an impressive accuracy of 98.79% in recognizing porcine cough. These results further underscore the effectiveness of combining acoustic source features with heterogeneous deep thermal source features, thereby establishing a robust feature representation for porcine cough recognition.

Sow serenity: automatic long-term measurement of lying behavior in crates and free-farrowing pens using 3D accelerometers

Accelerometers are useful in analyzing lying behavior in farm animals. The effect of the farrowing system on sow lying behavior has been studied around parturition, but not long-term. In a natural environment, sows increase activity 14 d post parturition, which we expected to be also evident in housed sows when they can move freely. The objective of this study was (1) to validate the methodology to automatically measure sow lying bouts and duration with accelerometers and (2) to apply it to crated and free-farrowing sows 24-h pre-parturition until weaning. We used videos with manual behavior coding as the gold standard for validation and calculated the agreement with an intraclass correlation coefficient (ICC), which was 0.30 (95% CI: -0.10 to 0.64) for the number of lying bouts. When transitional sitting bouts were excluded from the video dataset, the ICC for lying bouts increased to 0.86 (95% CI: 0.40 to 0.95). For lying duration, the ICC was 0.93 (95% CI: 0.26 to 0.98). We evaluated the effects of housing, day relative to parturition, and time of day on lying using the accelerometer data and linear mixed models. In crated sows, the number of lying bouts increased toward parturition, peaking at about five bouts per 6 h, and decreased to almost zero bouts after parturition. Then, it increased again (P = 0.001). In free-farrowing sows, the number of lying bouts gradually decreased from a high level towards parturition and was lowest after parturition. It remained constant, as in the crated sows, until day 15, when the number of bouts increased to eight bouts on day 20 (P = 0.001). Sows in both systems were lying almost all of the time between 18:00 and 00:00 hours and on all days (P = 0.001). The crated sows showed a very similar pattern in the other three-quarters of the day with a reduced lying time before parturition, a peak after parturition, reduced lying time from days 5 to 20, and an increase again towards weaning (P = 0.001). Free-farrowing sows had a similar pattern to the crated sows from 00:00 to 06:00 hours, but without the reduction in lying time from days 5 to 20. They showed an increase in lying time toward parturition, which remained constant with a final decrease toward weaning, especially during the day (P = 0.001). This study proves the accuracy of accelerometer-based sow lying behavior classification and shows that free-farrowing systems benefit lactating sows around parturition but also towards weaning in the nest-leaving phase by facilitating activity.

Thermal imaging and computer vision technologies for the enhancement of pig husbandry: a review

Pig farming, a vital industry, necessitates proactive measures for early disease detection and crush symptom monitoring to ensure optimum pig health and safety. This review explores advanced thermal sensing technologies and computer vision-based thermal imaging techniques employed for pig disease and piglet crush symptom monitoring on pig farms. Infrared thermography (IRT) is a non-invasive and efficient technology for measuring pig body temperature, providing advantages such as non-destructive, long-distance, and high-sensitivity measurements. Unlike traditional methods, IRT offers a quick and labor-saving approach to acquiring physiological data impacted by environmental temperature, crucial for understanding pig body physiology and metabolism. IRT aids in early disease detection, respiratory health monitoring, and evaluating vaccination effectiveness. Challenges include body surface emissivity variations affecting measurement accuracy. Thermal imaging and deep learning algorithms are used for pig behavior recognition, with the dorsal plane effective for stress detection. Remote health monitoring through thermal imaging, deep learning, and wearable devices facilitates non-invasive assessment of pig health, minimizing medication use. Integration of advanced sensors, thermal imaging, and deep learning shows potential for disease detection and improvement in pig farming, but challenges and ethical considerations must be addressed for successful implementation. This review summarizes the state-of-the-art technologies used in the pig farming industry, including computer vision algorithms such as object detection, image segmentation, and deep learning techniques. It also discusses the benefits and limitations of IRT technology, providing an overview of the current research field. This study provides valuable insights for researchers and farmers regarding IRT application in pig production, highlighting notable approaches and the latest research findings in this field.

Estimation of gestating sows' welfare status based on machine learning methods and behavioral data

Estimating the welfare status at an individual level on the farm is a current issue to improve livestock animal monitoring. New technologies showed opportunities to analyze livestock behavior with machine learning and sensors. The aim of the study was to estimate some components of the welfare status of gestating sows based on machine learning methods and behavioral data. The dataset used was a combination of individual and group measures of behavior (activity, social and feeding behaviors). A clustering method was used to estimate the welfare status of 69 sows (housed in four groups) during different periods (sum of 2 days per week) of gestation (between 6 and 10 periods, depending on the group). Three clusters were identified and labelled (scapegoat, gentle and aggressive). Environmental conditions and the sows' health influenced the proportion of sows in each cluster, contrary to the characteristics of the sow (age, body weight or body condition). The results also confirmed the importance of group behavior on the welfare of each individual. A decision tree was learned and used to classify the sows into the three categories of welfare issued from the clustering step. This classification relied on data obtained from an automatic feeder and automated video analysis, achieving an accuracy rate exceeding 72%. This study showed the potential of an automatic decision support system to categorize welfare based on the behavior of each gestating sow and the group of sows.

Three-dimensional surface motion capture of multiple freely moving pigs using MAMMAL

Understandings of the three-dimensional social behaviors of freely moving large-size mammals are valuable for both agriculture and life science, yet challenging due to occlusions in close interactions. Although existing animal pose estimation methods captured keypoint trajectories, they ignored deformable surfaces which contained geometric information essential for social interaction prediction and for dealing with the occlusions. In this study, we develop a Multi-Animal Mesh Model Alignment (MAMMAL) system based on an articulated surface mesh model. Our self-designed MAMMAL algorithms automatically enable us to align multi-view images into our mesh model and to capture 3D surface motions of multiple animals, which display better performance upon severe occlusions compared to traditional triangulation and allow complex social analysis. By utilizing MAMMAL, we are able to quantitatively analyze the locomotion, postures, animal-scene interactions, social interactions, as well as detailed tail motions of pigs. Furthermore, experiments on mouse and Beagle dogs demonstrate the generalizability of MAMMAL across different environments and mammal species.

Modelling of Animal Activity, Illuminance, and Noise on a Weaned Piglet Farm

Measuring animal activity and its evolution in real time is useful for animal welfare assessment. In addition, illuminance and noise level are two factors that can improve our understanding of animal activity. This study aims to establish relationships between animal activity as measured by passive infrared sensors, and both illuminance and noise level on a conventional weaned piglet farm. First, regression models were applied, and then cosine models with three harmonics were developed using least squares with a Generalized Reduced Gradient Nonlinear method. Finally, all the models were validated. Linear models showed positive correlations, with values between 0.40 and 0.56. Cosine models drew clear patterns of daily animal activity, illuminance and noise level with two peaks, one in the morning and one in the afternoon, coinciding with human activity inside the building, with a preference for inactivity at night-time and around midday. Cosine model fitting revealed strong correlations, both in the measurement and validation periods, for animal activity (R = 0.97 and 0.92), illuminance (R = 0.95 and 0.91) and noise level (R = 0.99 and 0.92). The developed models could be easily implemented in animal welfare monitoring systems and could provide useful information about animal activity through continuous monitoring of illuminance or noise levels.

Evaluating the Activity of Pigs with Radio-Frequency Identification and Virtual Walking Distances

Monitoring the activity of animals can help with assessing their health status. We monitored the walking activity of fattening pigs using a UHF-RFID system. Four hundred fattening pigs with UHF-RFID ear tags were recorded by RFID antennas at the troughs, playing devices and drinkers during the fattening period. A minimum walking distance, or virtual walking distance, was determined for each pig per day by calculating the distances between two consecutive reading areas. This automatically calculated value was used as an activity measure and not only showed differences between the pigs but also between different fattening stages. The longer the fattening periods lasted, the less walking activity was detected. The virtual walking distance ranged between 281 m on average in the first fattening stage and about 141 m in the last fattening stage in a restricted environment. The findings are similar to other studies considering walking distances of fattening pigs, but are far less labor-intensive and time-consuming than direct observations.

Scientific Productions on Precision Livestock Farming: An Overview of the Evolution and Current State of Research Based on a Bibliometric Analysis

The interest in precision livestock farming (PLF)-a concept discussed for the first time in the early 2000s-has advanced considerably in recent years due to its important role in the development of sustainable livestock production systems. However, a comprehensive bibliometric analysis of the PLF literature is lacking. To address this gap, this study analyzed documents published from 2005 to 2021, aiming to understand the historical influences on technology adoption in livestock farming, identify future global trends, and examine shifts in scientific research on this topic. By using specific search terms in the Web of Science Core Collection, 886 publications were identified and analyzed using the bibliometrix R-package. The analysis revealed that the collection consisted mostly of research articles (74.6%) and reviews (10.4%). The top three core journals were the Journal of Dairy Science, Computers and Electronics in Agriculture, and Animals. Over time, the number of publications has steadily increased, with a higher growth rate in the last five years (29.0%) compared to the initial period (13.7%). Authors and institutions from multiple countries have contributed to the literature, with the USA, the Netherlands, and Italy leading in terms of publication numbers. The analysis also highlighted the growing interest in bovine production systems, emphasizing the importance of behavioral studies in PLF tool development. Automated milking systems were identified as central drivers of innovation in the PLF sector. Emerging themes for the future included "emissions" and "mitigation", indicating a focus on environmental concerns.

Review: Fundamentals, limitations and pitfalls on the development and application of precision nutrition techniques for precision livestock farming

Precision livestock farming (PLF) concerns the management of livestock using the principles and technologies of process engineering. Precision nutrition (PN) is part of the PLF approach and involves the use of feeding techniques that allow the proper amount of feed with the suitable composition to be supplied in a timely manner to individual animals or groups of animals. Automatic data collection, data processing, and control actions are required activities for PN applications. Despite the benefits that PN offers to producers, few systems have been successfully implemented so far. Besides the economical and logistical challenges, there are conceptual limitations and pitfalls that threaten the widespread adoption of PN. Developers have to avoid the temptation of looking for the application of available sensors and instead concentrate on identifying the most appropriate and relevant information needed for the optimal functioning of PN applications. Efficient PN applications are obtained by controlling the nutrient requirement variations occurring between animals and over time. The utilization of feedback control algorithms for the automatic determination of optimal nutrient supply is not recommended. Mathematical models are the preferred data processing method for PN, but these models have to be designed to operate in real time using up-to-date information. These models are therefore structurally different than traditional nutrition or growth models. Combining knowledge- and data-driven models using machine learning and deep learning algorithms will enhance our ability to use real-time farm data, thus opening up new opportunities for PN. To facilitate the implementation of PN in farms, different experts and stakeholders should be involved in the development of the fully integrated and automatic PLF system. Precision livestock farming and PN should not be seen as just being a question of technology, but a successful marriage between knowledge and technology.

Grazing Sheep Behaviour Recognition Based on Improved YOLOV5

Fundamental sheep behaviours, for instance, walking, standing, and lying, can be closely associated with their physiological health. However, monitoring sheep in grazing land is complex as limited range, varied weather, and diverse outdoor lighting conditions, with the need to accurately recognise sheep behaviour in free range situations, are critical problems that must be addressed. This study proposes an enhanced sheep behaviour recognition algorithm based on the You Only Look Once Version 5 (YOLOV5) model. The algorithm investigates the effect of different shooting methodologies on sheep behaviour recognition and the model's generalisation ability under different environmental conditions and, at the same time, provides an overview of the design for the real-time recognition system. The initial stage of the research involves the construction of sheep behaviour datasets using two shooting methods. Subsequently, the YOLOV5 model was executed, resulting in better performance on the corresponding datasets, with an average accuracy of over 90% for the three classifications. Next, cross-validation was employed to verify the model's generalisation ability, and the results indicated the handheld camera-trained model had better generalisation ability. Furthermore, the enhanced YOLOV5 model with the addition of an attention mechanism module before feature extraction results displayed a mAP_@0.5 of 91.8% which represented an increase of 1.7%. Lastly, a cloud-based structure was proposed with the Real-Time Messaging Protocol (RTMP) to push the video stream for real-time behaviour recognition to apply the model in a practical situation. Conclusively, this study proposes an improved YOLOV5 algorithm for sheep behaviour recognition in pasture scenarios. The model can effectively detect sheep's daily behaviour for precision livestock management, promoting modern husbandry development.

Are biters sick? Health status of tail biters in comparison to control pigs

BACKGROUND

Tail biting is a multifactorial problem. As the health status is one of the factors commonly linked to tail biting, this study focuses on the health of identified biters. 30 (obsessive) biters are compared to 30 control animals by clinical and pathological examination as well as blood and cerebrospinal fluid samples. In that way, altogether 174 variables are compared between the groups. Moreover, connections between the variables are analysed.

RESULTS

In the clinical examination, 6 biters, but only 2 controls (P = 0.019) were noticeably agitated in the evaluation of general behaviour, while 8 controls were noticeably calmer (2 biters, P = 0.02). Biters had a lower body weight (P = 0.0007) and 13 biters had overlong bristles (4 controls, P = 0.008). In the pathological examination, 5 biters, but none of the controls had a hyperceratosis or inflammation of the pars proventricularis of the stomach (P = 0.018). However, 7 controls and only 3 biters were affected by gut inflammation (P = 0.03). In the blood sample, protein and albumin levels were below normal range for biters (protein: 51.6 g/l, albumin: 25.4 g/l), but not for controls (protein: 53.7 g/l, albumin: 27.4 g/l), (protein: P = 0.05, albumin: P = 0.02). Moreover, 14 biters, but only 8 controls had poikilocytosis (P = 0.05). Although not statistically different between groups, many animals (36/60) were affected by hypoproteinemia and hyponatremia as well as by hypokalemia (53/60) and almost all animals (58/60) had hypomagnesemia. For hypomagnesemia, significant connections with variables linked to tail damage and ear necrosis were detected (r_s/V/ρ ≥ 0.4, P ≤ 0.05).

CONCLUSION

The results suggest that behavioural tests might be helpful in identifying biters. Moreover, cornification and inflammation of the pars proventricularis is linked to becoming a biter. Furthermore, the results highlight the need for appropriate and adjusted nutrient and mineral supply, especially with regard to magnesium.

A Pig Mass Estimation Model Based on Deep Learning without Constraint

The body mass of pigs is an essential indicator of their growth and health. Lately, contactless pig body mass estimation methods based on computer vision technology have gained attention thanks to their potential to improve animal welfare and ensure breeders' safety. Nonetheless, current methods require pigs to be restrained in a confinement pen, and no study has been conducted in an unconstrained environment. In this study, we develop a pig mass estimation model based on deep learning, capable of estimating body mass without constraints. Our model comprises a Mask R-CNN-based pig instance segmentation algorithm, a Keypoint R-CNN-based pig keypoint detection algorithm and an improved ResNet-based pig mass estimation algorithm that includes multi-branch convolution, depthwise convolution, and an inverted bottleneck to improve accuracy. We constructed a dataset for this study using images and body mass data from 117 pigs. Our model achieved an RMSE of 3.52 kg on the test set, which is lower than that of the pig body mass estimation algorithm with ResNet and ConvNeXt as the backbone network, and the average estimation speed was 0.339 s·frame^-1 Our model can evaluate the body quality of pigs in real-time to provide data support for grading and adjusting breeding plans, and has broad application prospects.

Toward the automated detection of behavioral changes associated with the post-weaning transition in pigs

We modified an automated method capable of quantifying behaviors which we then applied to the changes associated with the post-weaning transition in pigs. The method is data-driven and depends solely on video-captured image data without relying on sensors or additional pig markings. It was applied to video images generated from an experiment during which post-weaned piglets were subjected to treatments either containing or not containing in-feed antimicrobials (ZnO or antibiotics). These treatments were expected to affect piglet performance and health in the short-term by minimizing the risk from post-weaning enteric disorders, such as diarrhea. The method quantified total group feeding and drinking behaviors as well as posture (i.e., standing and non-standing) during the first week post-weaning, when the risk of post-weaning diarrhea is at its highest, by learning from the variations within each behavior using data manually annotated by a behavioral scientist. Automatically quantified changes in behavior were consistent with the effects of the absence of antimicrobials on pig performance and health, and manifested as reduced feed efficiency and looser feces. In these piglets both drinking and standing behaviors were increased during the first 6 days post-weaning. The correlation between fecal consistency and drinking behavior 6 days post weaning was relatively high, suggesting that these behaviors may have a diagnostic value. The presence or absence of in-feed antimicrobials had no effect on feeding behavior, which, however, increased over time. The approach developed here is capable of automatically monitoring several different behaviors of a group of pigs at the same time, and potentially this may be where its value as a diagnostic tool may lie.

Identifying Early Indicators of Tail Biting in Pigs by Variable Selection Using Partial Least Squares Regression

This study examined relevant variables for predicting the prevalence of pigs with a tail lesion in rearing (REA) and fattening (FAT). Tail lesions were recorded at two scoring days a week in six pens in both REA (10 batches, 840 scoring days) and FAT (5 batches, 624 scoring days). To select the variables that best explain the variation within the prevalence of pigs with a tail lesion, partial least squares regression models were used with the variable importance in projection (VIP) and regression coefficients (β) as selection criteria. In REA, five factors were extracted explaining 60.6% of the dependent variable's variance, whereas in FAT five extracted factors explained 62.4% of the dependent variable's variance. According to VIP and β, seven variables were selected in REA and six in FAT with the tail posture being the most important variable. In addition, skin lesions, treatment index in the suckling phase, water consumption (mean), activity time (mean; CV) and exhaust air rate (CV) were selected in REA. In FAT, additional musculoskeletal system issues, activity time (mean; CV) and exhaust air rate (mean; CV) were selected according to VIP and β. The selected variables indicate which variables should be collected in the stable to e.g., predict tail biting.

Welfare of pigs on farm

This scientific opinion focuses on the welfare of pigs on farm, and is based on literature and expert opinion. All pig categories were assessed: gilts and dry sows, farrowing and lactating sows, suckling piglets, weaners, rearing pigs and boars. The most relevant husbandry systems used in Europe are described. For each system, highly relevant welfare consequences were identified, as well as related animal-based measures (ABMs), and hazards leading to the welfare consequences. Moreover, measures to prevent or correct the hazards and/or mitigate the welfare consequences are recommended. Recommendations are also provided on quantitative or qualitative criteria to answer specific questions on the welfare of pigs related to tail biting and related to the European Citizen's Initiative 'End the Cage Age'. For example, the AHAW Panel recommends how to mitigate group stress when dry sows and gilts are grouped immediately after weaning or in early pregnancy. Results of a comparative qualitative assessment suggested that long-stemmed or long-cut straw, hay or haylage is the most suitable material for nest-building. A period of time will be needed for staff and animals to adapt to housing lactating sows and their piglets in farrowing pens (as opposed to crates) before achieving stable welfare outcomes. The panel recommends a minimum available space to the lactating sow to ensure piglet welfare (measured by live-born piglet mortality). Among the main risk factors for tail biting are space allowance, types of flooring, air quality, health status and diet composition, while weaning age was not associated directly with tail biting in later life. The relationship between the availability of space and growth rate, lying behaviour and tail biting in rearing pigs is quantified and presented. Finally, the panel suggests a set of ABMs to use at slaughter for monitoring on-farm welfare of cull sows and rearing pigs.

Water consumption and wastage behaviour in pigs: implications for antimicrobial administration and stewardship

Daily water use and wastage patterns of pigs have major effects on the efficacy of in-water antimicrobial dosing events when conducted for metaphylaxis or to treat clinical disease. However, daily water use and wastage patterns of pigs are not routinely quantified on farms and are not well understood. We conducted a prospective, observational 27-day study of the daily water use and wastage patterns of a pen group of 15 finisher pigs reared in a farm building. We found that the group of pigs wasted a median of 36.5% of the water used per day. We developed models of the patterns of water used and wasted by pigs over each 24-h period using a Bayesian statistical method with the brm() function in the brms package. Both patterns were uni-modal, peaking at 1400-1700, and closely aligned. Wastage was slightly greater during hours of higher water use. We have shown that it is feasible to quantify the water use and wastage patterns of pigs in farm buildings using a system that records and aggregates data, and analyses them using hierarchical generalised additive models. This system could support more efficacious in-water antimicrobial dosing on farms, and better antimicrobial stewardship, by helping to reduce the quantities of antimicrobials used and disseminated into the environment.

A Review of Monitoring Techniques for Livestock Respiration and Sounds

This article reviews the different techniques used to monitor the respiration and sounds of livestock. Livestock respiration is commonly assessed visually by observing abdomen fluctuation; however, the traditional methods are time consuming, subjective, being therefore impractical for large-scale operations and must rely on automation. Contact and non-contact technologies are used to automatically monitor respiration rate; contact technologies (e.g., accelerometers, pressure sensors, and thermistors) utilize sensors that are physically mounted on livestock while non-contact technologies (e.g., computer vision, thermography, and sound analysis) enable a non-invasive method of monitoring respiration. This work summarizes the advantages and disadvantages of contact and non-contact technologies and discusses the emerging role of non-contact sensors in automating monitoring for large-scale farming operations. This work is the first in-depth examination of automated monitoring technologies for livestock respiratory diseases; the findings and recommendations are important for livestock researchers and practitioners who can gain a better understanding of these different technologies, especially emerging non-contact sensing.

Personality and predictability in farmed calves using movement and space-use behaviours quantified by ultra-wideband sensors

Individuals within a population often show consistent between individual differences in their average behavioural expression (personality), and consistent differences in their within individual variability of behaviour around the mean (predictability). Where correlations between different personality traits and/or the predictability of traits exist, these represent behavioural or predictability syndromes. In wild populations, behavioural syndromes have consequences for individuals' survival and reproduction and affect the structure and functioning of groups and populations. The consequences of behavioural syndromes for farm animals are less well explored, partly due to the challenges in quantifying behaviour of many individuals across time and context in a farm setting. Here, we use ultra-wideband location sensors to provide precise measures of movement and space use for 60 calves over 40-48 days. We are the first livestock study to demonstrate consistent within and between individual variation in movement and space use with repeatability values of up to 0.80 and CVp values up to 0.49. Our results show correlations in personality and predictability, indicating the existence of 'exploratory' and 'active' personality traits in farmed calves. We consider the consequences of such individual variability for cattle behaviour and welfare and how such data may be used to inform management decisions in farm animals.

Impact of housing environment and management on pre-/post-weaning piglet productivity

The complex environment surrounding young pigs reared in intensive housing systems directly influences their productivity and livelihood. Much of the seminal literature utilized housing and husbandry practices that have since drastically evolved through advances in genetic potential, nutrition, health, and technology. This review focuses on the environmental interaction and responses of pigs during the first 8 wk of life, separated into pre-weaning (creep areas) and post-weaning (nursery or wean-finish) phases. Further, a perspective on instrumentation and precision technologies for animal-based (physiological and behavioral) and environmental measures documents current approaches and future possibilities. A warm microclimate for piglets during the early days of life, especially the first 12 h, is critical. While caretaker interventions can mitigate the extent of hypothermia, low birth weight remains a dominant risk factor for mortality. Post-weaning, the thermoregulation capabilities have improved, but subsequent transportation, nutritional, and social stressors enhance the requisite need for a warm, low draft environment with the proper flooring. A better understanding of the individual environmental factors that affect young pigs as well as the creation of comprehensive environment indices or improved, non-contact sensing technology is needed to better evaluate and manage piglet environments. Such enhanced understanding and evaluation of pig–environment interaction could lead to innovative environmental control and husbandry interventions to foster healthy and productive pigs.

GAN-Based Video Denoising with Attention Mechanism for Field-Applicable Pig Detection System

Infrared cameras allow non-invasive and 24 h continuous monitoring. Thus, they are widely used in automatic pig monitoring, which is essential to maintain the profitability and sustainability of intensive pig farms. However, in practice, impurities such as insect secretions continuously pollute camera lenses. This causes problems with IR reflections, which can seriously affect pig detection performance. In this study, we propose a noise-robust, real-time pig detection system that can improve accuracy in pig farms where infrared cameras suffer from the IR reflection problem. The system consists of a data collector to collect infrared images, a preprocessor to transform noisy images into clean images, and a detector to detect pigs. The preprocessor embeds a multi-scale spatial attention module in U-net and generative adversarial network (GAN) models, enabling the model to pay more attention to the noisy area. The GAN model was trained on paired sets of clean data and data with simulated noise. It can operate in a real-time and end-to-end manner. Experimental results show that the proposed preprocessor was able to significantly improve the average precision of pig detection from 0.766 to 0.906, with an additional execution time of only 4.8 ms on a PC environment.

Visual Sensor Placement Optimization with 3D Animation for Cattle Health Monitoring in a Confined Operation

Simple Summary

This paper introduces a new method of finding the best locations to place video cameras inside large cattle barns to monitor the behavior and health of the animals. Current approaches to livestock video monitoring rely on mounting cameras in the most convenient places for installation, but those locations might either be impractical for actual barns and/or might not capture the best views. This work showed that there is short list of the best placement options for the cameras to choose from which will provide the best camera views.

Abstract

Computer vision has been extensively used for livestock welfare monitoring in recent years, and data collection with a sensor or camera is the first part of the complete workflow. While current practice in computer vision-based animal welfare monitoring often analyzes data collected from a sensor or camera mounted on the roof or ceiling of a laboratory, such camera placement is not always viable in a commercial confined cattle feeding environment. This study therefore sought to determine the optimal camera placement locations in a confined steer feeding operation. Measurements of cattle pens were used to create a 3D farm model using Blender 3D computer graphic software. In the first part of this study, a method was developed to calculate the camera coverage in a 3D farm environment, and in the next stage, a genetic algorithm-based model was designed for finding optimal placements of a multi-camera and multi-pen setup. The algorithm’s objective was to maximize the multi-camera coverage while minimizing budget. Two different optimization methods involving multiple cameras and pen combinations were used. The results demonstrated the applicability of the genetic algorithm in achieving the maximum coverage and thereby enhancing the quality of the livestock visual-sensing data. The algorithm also provided the top 25 solutions for each camera and pen combination with a maximum coverage difference of less than 3.5% between them, offering numerous options for the farm manager.

Use of Host Feeding Behavior and Gut Microbiome Data in Estimating Variance Components and Predicting Growth and Body Composition Traits in Swine

The purpose of this study was to investigate the use of feeding behavior in conjunction with gut microbiome sampled at two growth stages in predicting growth and body composition traits of finishing pigs. Six hundred and fifty-one purebred boars of three breeds: Duroc (DR), Landrace (LR), and Large White (LW), were studied. Feeding activities were recorded individually from 99 to 163 days of age. The 16S rRNA gene sequences were obtained from each pig at 123 ± 4 and 158 ± 4 days of age. When pigs reached market weight, body weight (BW), ultrasound backfat thickness (BF), ultrasound loin depth (LD), and ultrasound intramuscular fat (IMF) content were measured on live animals. Three models including feeding behavior (Model_FB), gut microbiota (Model_M), or both (Model_FB_M) as predictors, were investigated. Prediction accuracies were evaluated through cross-validation across genetic backgrounds using the leave-one-breed-out strategy and across rearing environments using the leave-one-room-out approach. The proportions of phenotypic variance of growth and body composition traits explained by feeding behavior ranged from 0.02 to 0.30, and from 0.20 to 0.52 when using gut microbiota composition. Overall prediction accuracy (averaged over traits and time points) of phenotypes was 0.24 and 0.33 for Model_FB, 0.27 and 0.19 for Model_M, and 0.40 and 0.35 for Model_FB_M for the across-breed and across-room scenarios, respectively. This study shows how feeding behavior and gut microbiota composition provide non-redundant information in predicting growth in swine.

Relevance, Impartiality, Welfare and Consent: Principles of an Animal-Centered Research Ethics

The principles of Replacement, Reduction and Refinement (3Rs) were developed to address the ethical dilemma that arises from the use of animals, without their consent, in procedures that may harm them but that are deemed necessary to achieve a greater good. While aiming to protect animals, the 3Rs are underpinned by a process-centered ethical perspective which regards them as instruments in a scientific apparatus. This paper explores the applicability of an animal-centered ethics to animal research, whereby animals would be regarded as autonomous subjects, legitimate stakeholders in and contributors to a research process, with their own interests and capable of consenting and dissenting to their involvement. This perspective derives from the ethical stance taken within the field of Animal-Computer Interaction (ACI), where researchers acknowledge that an animal-centered approach is essential to ensuring the best research outcomes. We propose the ethical principles of relevance, impartiality, welfare and consent, and a scoring system to help researchers and delegated authorities assess the extent to which a research procedure aligns with them. This could help researchers determine when being involved in research is indeed in an animal's best interests, when a procedure could be adjusted to increase its ethical standard or when the use of non-animal methods is more urgently advisable. We argue that the proposed principles should complement the 3Rs within an integrated ethical framework that recognizes animals' autonomy, interests and role, for a more nuanced ethical approach and for supporting the best possible research for the benefit animal partakers and wider society.

EmbeddedPigCount: Pig Counting with Video Object Detection and Tracking on an Embedded Board

Knowing the number of pigs on a large-scale pig farm is an important issue for efficient farm management. However, counting the number of pigs accurately is difficult for humans because pigs do not obediently stop or slow down for counting. In this study, we propose a camera-based automatic method to count the number of pigs passing through a counting zone. That is, using a camera in a hallway, our deep-learning-based video object detection and tracking method analyzes video streams and counts the number of pigs passing through the counting zone. Furthermore, to execute the counting method in real time on a low-cost embedded board, we consider the tradeoff between accuracy and execution time, which has not yet been reported for pig counting. Our experimental results on an NVIDIA Jetson Nano embedded board show that this “light-weight” method is effective for counting the passing-through pigs, in terms of both accuracy (i.e., 99.44%) and execution time (i.e., real-time execution), even when some pigs pass through the counting zone back and forth.

Repeatability and Predictability of Calf Feeding Behaviors-Quantifying Between- and Within-Individual Variation for Precision Livestock Farming

Individual calves show substantial between- and within-individual variation in their feeding behavior, the existence and extent of which are not fully researched. In this study, 57,196 feeding records, collected by a computerized milk feeder from 48 pre-weaned calves over 5 weeks, were collated and analyzed for individual differences in three different feeding behaviors using a multi-level modeling approach. For each feeding behavior, we quantified behavioral variation by calculating repeatability and the coefficient of variation in predictability. Our results indicate that calves differed from each other in their average behavioral expression (behavioral type) and in their residual, within individual variation around their behavioral type (predictability). Feeding rate and total meals had the highest repeatability (>0.4) indicating that substantial, temporally stable between-individual differences exist for these behaviors. Additionally, for some behaviors (e.g., feeding rate) calves varied from more to less predictable whereas for other behaviors (e.g., meal size) calves were more homogenous in their within-individual variation around their behavioral type. Finally, we show that for individual calves, behavioral types for feeding rate and total meals were positively correlated which may suggest the existence of an underlying factor responsible for driving the (co)expression of these two behaviors. Our results highlight how the application of methods from the behavioral ecology literature can assist in improving our understanding of individual differences in calf feeding behavior. Furthermore, by uncovering consistencies between individual behavioral differences in calves, our results indicate that animal personality may play a role in driving variability in calf feeding behavior.

Applications of Smart Technology as a Sustainable Strategy in Modern Swine Farming

The size of the pork market is increasing globally to meet the demand for animal protein, resulting in greater farm size for swine and creating a great challenge to swine farmers and industry owners in monitoring the farm activities and the health and behavior of the herd of swine. In addition, the growth of swine production is resulting in a changing climate pattern along with the environment, animal welfare, and human health issues, such as antimicrobial resistance, zoonosis, etc. The profit of swine farms depends on the optimum growth and good health of swine, while modern farming practices can ensure healthy swine production. To solve these issues, a future strategy should be considered with information and communication technology (ICT)-based smart swine farming, considering auto-identification, remote monitoring, feeding behavior, animal rights/welfare, zoonotic diseases, nutrition and food quality, labor management, farm operations, etc., with a view to improving meat production from the swine industry. Presently, swine farming is not only focused on the development of infrastructure but is also occupied with the application of technological knowledge for designing feeding programs, monitoring health and welfare, and the reproduction of the herd. ICT-based smart technologies, including smart ear tags, smart sensors, the Internet of Things (IoT), deep learning, big data, and robotics systems, can take part directly in the operation of farm activities, and have been proven to be effective tools for collecting, processing, and analyzing data from farms. In this review, which considers the beneficial role of smart technologies in swine farming, we suggest that smart technologies should be applied in the swine industry. Thus, the future swine industry should be automated, considering sustainability and productivity.

Computer Vision for Detection of Body Posture and Behavior of Red Foxes

The behavior of animals is related to their health and welfare status. The latter plays a particular role in animal experiments, where continuous monitoring is essential for animal welfare. In this study, we focus on red foxes in an experimental setting and study their behavior. Although animal behavior is a complex concept, it can be described as a combination of body posture and activity. To measure body posture and activity, video monitoring can be used as a non-invasive and cost-efficient tool. While it is possible to analyze the video data resulting from the experiment manually, this method is time consuming and costly. We therefore use computer vision to detect and track the animals over several days. The detector is based on a neural network architecture. It is trained to detect red foxes and their body postures, i.e., ‘lying’, ‘sitting’, and ‘standing’. The trained algorithm has a mean average precision of 99.91%. The combination of activity and posture results in nearly continuous monitoring of animal behavior. Furthermore, the detector is suitable for real-time evaluation. In conclusion, evaluating the behavior of foxes in an experimental setting using computer vision is a powerful tool for cost-efficient real-time monitoring.

A Systematic Review of Automatic Health Monitoring in Calves: Glimpsing the Future From Current Practice

Infectious diseases, particularly bovine respiratory disease (BRD) and neonatal calf diarrhea (NCD), are prevalent in calves. Efficient health-monitoring tools to identify such diseases on time are lacking. Common practice (i.e., health checks) often identifies sick calves at a late stage of disease or not at all. Sensor technology enables the automatic and continuous monitoring of calf physiology or behavior, potentially offering timely and precise detection of sick calves. A systematic overview of automated disease detection in calves is still lacking. The objectives of this literature review were hence: to investigate previously applied sensor validation methods used in the context of calf health, to identify sensors used on calves, the parameters these sensors monitor, and the statistical tools applied to identify diseases, to explore potential research gaps and to point to future research opportunities. To achieve these objectives, systematic literature searches were conducted. We defined four stages in the development of health-monitoring systems: (1) sensor technique, (2) data interpretation, (3) information integration, and (4) decision support. Fifty-four articles were included (stage one: 26; stage two: 19; stage three: 9; and stage four: 0). Common parameters that assess the performance of these systems are sensitivity, specificity, accuracy, precision, and negative predictive value. Gold standards that typically assess these parameters include manual measurement and manual health-assessment protocols. At stage one, automatic feeding stations, accelerometers, infrared thermography cameras, microphones, and 3-D cameras are accurate in screening behavior and physiology in calves. At stage two, changes in feeding behaviors, lying, activity, or body temperature corresponded to changes in health status, and point to health issues earlier than manual health checks. At stage three, accelerometers, thermometers, and automatic feeding stations have been integrated into one system that was shown to be able to successfully detect diseases in calves, including BRD and NCD. We discuss these findings, look into potentials at stage four, and touch upon the topic of resilience, whereby health-monitoring system might be used to detect low resilience (i.e., prone to disease but clinically healthy calves), promoting further improvements in calf health and welfare.

Detecting Animal Contacts-A Deep Learning-Based Pig Detection and Tracking Approach for the Quantification of Social Contacts

The identification of social interactions is of fundamental importance for animal behavioral studies, addressing numerous problems like investigating the influence of social hierarchical structures or the drivers of agonistic behavioral disorders. However, the majority of previous studies often rely on manual determination of the number and types of social encounters by direct observation which requires a large amount of personnel and economical efforts. To overcome this limitation and increase research efficiency and, thus, contribute to animal welfare in the long term, we propose in this study a framework for the automated identification of social contacts. In this framework, we apply a convolutional neural network (CNN) to detect the location and orientation of pigs within a video and track their movement trajectories over a period of time using a Kalman filter (KF) algorithm. Based on the tracking information, we automatically identify social contacts in the form of head–head and head–tail contacts. Moreover, by using the individual animal IDs, we construct a network of social contacts as the final output. We evaluated the performance of our framework based on two distinct test sets for pig detection and tracking. Consequently, we achieved a Sensitivity, Precision, and F1-score of 94.2%, 95.4%, and 95.1%, respectively, and a MOTA score of 94.4%. The findings of this study demonstrate the effectiveness of our keypoint-based tracking-by-detection strategy and can be applied to enhance animal monitoring systems.

Control of sheep flystrike: what's been tried in the past and where to from here

Flystrike remains a serious financial and animal welfare issue for the sheep industry in Australia despite many years of research into control methods. The present paper provides an extensive review of past research on flystrike, and highlights areas that hold promise for providing long-term control options. We describe areas where the application of modern scientific advances may provide increased impetus to some novel, as well as some previously explored, control methods. We provide recommendations for research activities: insecticide resistance management, novel delivery methods for therapeutics, improved breeding indices for flystrike-related traits, mechanism of nematode-induced scouring in mature animals. We also identify areas where advances can be made in flystrike control through the greater adoption of well-recognised existing management approaches: optimal insecticide-use patterns, increased use of flystrike-related Australian Sheep Breeding Values, and management practices to prevent scouring in young sheep. We indicate that breeding efforts should be primarily focussed on the adoption and improvement of currently available breeding tools and towards the future integration of genomic selection methods. We describe factors that will impact on the ongoing availability of insecticides for flystrike control and on the feasibility of vaccination. We also describe areas where the blowfly genome may be useful in providing impetus to some flystrike control strategies, such as area-wide approaches that seek to directly suppress or eradicate sheep blowfly populations. However, we also highlight the fact that commercial and feasibility considerations will act to temper the potential for the genome to act as the basis for providing some control options.

Deep-Learning-Based Automatic Monitoring of Pigs' Physico-Temporal Activities at Different Greenhouse Gas Concentrations

Pig behavior is an integral part of health and welfare management, as pigs usually reflect their inner emotions through behavior change. The livestock environment plays a key role in pigs' health and wellbeing. A poor farm environment increases the toxic GHGs, which might deteriorate pigs' health and welfare. In this study a computer-vision-based automatic monitoring and tracking model was proposed to detect pigs' short-term physical activities in the compromised environment. The ventilators of the livestock barn were closed for an hour, three times in a day (07:00-08:00, 13:00-14:00, and 20:00-21:00) to create a compromised environment, which increases the GHGs level significantly. The corresponding pig activities were observed before, during, and after an hour of the treatment. Two widely used object detection models (YOLOv4 and Faster R-CNN) were trained and compared their performances in terms of pig localization and posture detection. The YOLOv4, which outperformed the Faster R-CNN model, was coupled with a Deep-SORT tracking algorithm to detect and track the pig activities. The results revealed that the pigs became more inactive with the increase in GHG concentration, reducing their standing and walking activities. Moreover, the pigs shortened their sternal-lying posture, increasing the lateral lying posture duration at higher GHG concentration. The high detection accuracy (mAP: 98.67%) and tracking accuracy (MOTA: 93.86% and MOTP: 82.41%) signify the models' efficacy in the monitoring and tracking of pigs' physical activities non-invasively.

Intelligent Perception-Based Cattle Lameness Detection and Behaviour Recognition: A Review

Simple Summary

Cattle lameness detection as well as behaviour recognition are the two main objectives in the applications of precision livestock farming (PLF). Over the last five years, the development of smart sensors, big data, and artificial intelligence has offered more automatic tools. In this review, we discuss over 100 papers that used automated techniques to detect cattle lameness and to recognise animal behaviours. To assist researchers and policy-makers in promoting various livestock technologies for monitoring cattle welfare and productivity, we conducted a comprehensive investigation of intelligent perception for cattle lameness detection and behaviour analysis in the PLF domain. Based on the literature review, we anticipate that PLF will develop in an objective, autonomous, and real-time direction. Additionally, we suggest that further research should be dedicated to improving the data quality, modeling accuracy, and commercial availability.

Abstract

The growing world population has increased the demand for animal-sourced protein. However, animal farming productivity is faced with challenges from traditional farming practices, socioeconomic status, and climate change. In recent years, smart sensors, big data, and deep learning have been applied to animal welfare measurement and livestock farming applications, including behaviour recognition and health monitoring. In order to facilitate research in this area, this review summarises and analyses some main techniques used in smart livestock farming, focusing on those related to cattle lameness detection and behaviour recognition. In this study, more than 100 relevant papers on cattle lameness detection and behaviour recognition have been evaluated and discussed. Based on a review and a comparison of recent technologies and methods, we anticipate that intelligent perception for cattle behaviour and welfare monitoring will develop towards standardisation, a larger scale, and intelligence, combined with Internet of things (IoT) and deep learning technologies. In addition, the key challenges and opportunities of future research are also highlighted and discussed.

Early Detection of Diarrhea in Weaned Piglets From Individual Feed, Water and Weighing Data

This study analyzed individual water and feed consumption related to weight of weaned piglets and their link to diarrhea. Data were collected from 15 batches of 102 piglets each, using specific automata (connected feeders, connected drinkers, automatic weighing stations, RFID ear tags). Analyses were carried out every week on the 138 healthy animals compared by weight category. The average feed consumption had no significant difference between weight categories (light, medium, heavy pigs) whatever the week and was close to 4% of the live weight. For the average water consumption according to weight, it was close to 10%. There was no significant difference between weight groups, except at the end of the period, where the variability of one heavy pig was so high that its own water consumption caused significant difference when compared with the light group. But these overall stable averages do not highlight the high intra-individual variabilities, which was around 40% for both water and feed data at the beginning of trial. At the end, it was almost 16% for feed consumption and 25% for water. The comparison between healthy and diarrheic piglets showed no statistical difference for average water consumption on the day of the first clinical signs and even 1 and 2 days before. In contrast, the average feed consumption had a very significant difference (P ≤ 0.001) for days 5–7 after the weaning and a significant difference for day 8 (P ≤ 0.05). Differences were also significant for data 24 and 48 h before first clinical signs. This means either that diarrheic piglets decrease their feed consumption the first days after weaning or that it is because they eat less that they become diarrheic. So, the hypothesis was that feed consumption could be an interesting indicator to detect early sick animals. Nevertheless, despite this difference, machine learning methods failed in detecting individually diarrheic animals from water and feed consumption related to weight, because of considerable individual variability. To improve these results, one solution could be to collect other data from new sensors like automatic measurement of body temperature or location of piglets in the pen by image analysis.

A Systematic Review on Validated Precision Livestock Farming Technologies for Pig Production and Its Potential to Assess Animal Welfare

Several precision livestock farming (PLF) technologies, conceived for optimizing farming processes, are developed to detect the physical and behavioral changes of animals continuously and in real-time. The aim of this review was to explore the capacity of existing PLF technologies to contribute to the assessment of pig welfare. In a web search for commercially available PLF for pigs, 83 technologies were identified. A literature search was conducted, following systematic review guidelines (PRISMA), to identify studies on the validation of sensor technologies for assessing animal-based welfare indicators. Two validation levels were defined: internal (evaluation during system building within the same population that were used for system building) and external (evaluation on a different population than during system building). From 2,463 articles found, 111 were selected, which validated some PLF that could be applied to the assessment of animal-based welfare indicators of pigs (7% classified as external, and 93% as internal validation). From our list of commercially available PLF technologies, only 5% had been externally validated. The more often validated technologies were vision-based solutions (n = 45), followed by load-cells (n = 28; feeders and drinkers, force plates and scales), accelerometers (n = 14) and microphones (n = 14), thermal cameras (n = 10), photoelectric sensors (n = 5), radio-frequency identification (RFID) for tracking (n = 2), infrared thermometers (n = 1), and pyrometer (n = 1). Externally validated technologies were photoelectric sensors (n = 2), thermal cameras (n = 2), microphone (n = 1), load-cells (n = 1), RFID (n = 1), and pyrometer (n = 1). Measured traits included activity and posture-related behavior, feeding and drinking, other behavior, physical condition, and health. In conclusion, existing PLF technologies are potential tools for on-farm animal welfare assessment in pig production. However, validation studies are lacking for an important percentage of market available tools, and in particular research and development need to focus on identifying the feature candidates of the measures (e.g., deviations from diurnal pattern, threshold levels) that are valid signals of either negative or positive animal welfare. An important gap identified are the lack of technologies to assess affective states (both positive and negative states).

Pig Weight and Body Size Estimation Using a Multiple Output Regression Convolutional Neural Network: A Fast and Fully Automatic Method

Pig weight and body size are important indicators for producers. Due to the increasing scale of pig farms, it is increasingly difficult for farmers to quickly and automatically obtain pig weight and body size. Due to this problem, we focused on a multiple output regression convolutional neural network (CNN) to estimate pig weight and body size. DenseNet201, ResNet152 V2, Xception and MobileNet V2 were modified into multiple output regression CNNs and trained on modeling data. By comparing the estimated performance of each model on test data, modified Xception was selected as the optimal estimation model. Based on pig height, body shape, and contour, the mean absolute error (MAE) of the model to estimate body weight (BW), shoulder width (SW), shoulder height (SH), hip width (HW), hip width (HH), and body length (BL) were 1.16 kg, 0.33 cm, 1.23 cm, 0.38 cm, 0.66 cm, and 0.75 cm, respectively. The coefficient of determination (R²) value between the estimated and measured results was in the range of 0.9879–0.9973. Combined with the LabVIEW software development platform, this method can estimate pig weight and body size accurately, quickly, and automatically. This work contributes to the automatic management of pig farms.

Welfare Health and Productivity in Commercial Pig Herds

In recent years, there have been very dynamic changes in both pork production and pig breeding technology around the world. The general trend of increasing the efficiency of pig production, with reduced employment, requires optimisation and a comprehensive approach to herd management. One of the most important elements on the way to achieving this goal is to maintain animal welfare and health. The health of the pigs on the farm is also a key aspect in production economics. The need to maintain a high health status of pig herds by eliminating the frequency of different disease units and reducing the need for antimicrobial substances is part of a broadly understood high potential herd management strategy. Thanks to the use of sensors (cameras, microphones, accelerometers, or radio-frequency identification transponders), the images, sounds, movements, and vital signs of animals are combined through algorithms and analysed for non-invasive monitoring of animals, which allows for early detection of diseases, improves their welfare, and increases the productivity of breeding. Automated, innovative early warning systems based on continuous monitoring of specific physiological (e.g., body temperature) and behavioural parameters can provide an alternative to direct diagnosis and visual assessment by the veterinarian or the herd keeper.

A computer vision-based approach for behavior recognition of gestating sows fed different fiber levels during high ambient temperature

The objectives of this study were to evaluate convolutional neural network models and computer vision techniques for the classification of swine posture with high accuracy and to use the derived result in the investigation of the effect of dietary fiber level on the behavioral characteristics of the pregnant sow under low and high ambient temperatures during the last stage of gestation. A total of 27 crossbred sows (Yorkshire × Landrace; average body weight, 192.2 ± 4.8 kg) were assigned to three treatments in a randomized complete block design during the last stage of gestation (days 90 to 114). The sows in group 1 were fed a 3% fiber diet under neutral ambient temperature; the sows in group 2 were fed a diet with 3% fiber under high ambient temperature (HT); the sows in group 3 were fed a 6% fiber diet under HT. Eight popular deep learning-based feature extraction frameworks (DenseNet121, DenseNet201, InceptionResNetV2, InceptionV3, MobileNet, VGG16, VGG19, and Xception) used for automatic swine posture classification were selected and compared using the swine posture image dataset that was constructed under real swine farm conditions. The neural network models showed excellent performance on previously unseen data (ability to generalize). The DenseNet121 feature extractor achieved the best performance with 99.83% accuracy, and both DenseNet201 and MobileNet showed an accuracy of 99.77% for the classification of the image dataset. The behavior of sows classified by the DenseNet121 feature extractor showed that the HT in our study reduced (p < 0.05) the standing behavior of sows and also has a tendency to increase (p = 0.082) lying behavior. High dietary fiber treatment tended to increase (p = 0.064) lying and decrease (p < 0.05) the standing behavior of sows, but there was no change in sitting under HT conditions.

Feed intake and feeding behavior traits for gestating sows recorded using electronic sow feeders

Electronic sow feeding (ESF) systems are used to control feed delivery to individual sows that are group-housed. Feeding levels for gestating sows are typically restricted to prevent excessive body weight gain. Any alteration of intake from the allocated feeding curve or unusual feeding behavior could indicate potential health issues. The objective of this study was to use data recorded by ESF to establish and characterize novel feed intake and feeding behavior traits and to estimate their heritabilities. Raw data were available from two farms with in-house manufactured (Farm A) or commercial (Farm B) ESF. The traits derived included feed intake, time spent eating, and rate of feed consumption, averaged across or within specific time periods of gestation. Additional phenotypes included average daily number of feeding events (AFE), along with the cumulative numbers of days where sows spent longer than 30 min in the ESF (ABOVE30), missed their daily intake (MISSF), or consumed below 1 kg of feed (BELOW1). The appetite of sows was represented by averages of score (APPETITE), a binary value for allocation eaten or not (DA_bin), or the standard deviation of the difference between feed intake and allocation (SDA-I). Gilts took longer to eat than sows (15.5 ± 0.13 vs. 14.1 ± 0.11 min/d) despite a lower feed allocation (2.13 ± 0.00 vs. 2.36 ± 0.01 kg/d). The lowest heritability estimates (below 0.10) occurred for feed intake traits, due to the restriction in feed allocation, although heritabilities were slightly higher for Farm B, with restriction in the eating time. The low heritability for AFE (0.05 ± 0.02) may have reflected the lack of recording of nonfeeding visits, but repeatability was moderate (0.26 ± 0.03, Farm A). Time-related traits were moderately to highly heritable and repeatable, demonstrating genetic variation between individuals in their feeding behaviors. Heritabilities for BELOW1 (Farm A: 0.16 ± 0.04 and Farm B: 0.15 ± 0.09) and SDA-I (Farm A: 0.17 ± 0.04 and Farm B: 0.10 ± 0.08) were similar across farms. In contrast, MISSF was moderately heritable in Farm A (0.19 ± 0.04) but lowly heritable in Farm B (0.05 ± 0.07). Heritabilities for DA_bin were dissimilar between farms (Farm A: 0.02 ± 0.02 and Farm B: 0.23 ± 0.10) despite similar incidence. Individual phenotypes constructed from ESF data could be useful for genetic evaluation purposes, but equivalent capabilities to generate phenotypes were not available for both ESF systems.

Machine Learning Algorithms to Classify and Quantify Multiple Behaviours in Dairy Calves Using a Sensor: Moving beyond Classification in Precision Livestock

Previous research has shown that sensors monitoring lying behaviours and feeding can detect early signs of ill health in calves. There is evidence to suggest that monitoring change in a single behaviour might not be enough for disease prediction. In calves, multiple behaviours such as locomotor play, self-grooming, feeding and activity whilst lying are likely to be informative. However, these behaviours can occur rarely in the real world, which means simply counting behaviours based on the prediction of a classifier can lead to overestimation. Here, we equipped thirteen pre-weaned dairy calves with collar-mounted sensors and monitored their behaviour with video cameras. Behavioural observations were recorded and merged with sensor signals. Features were calculated for 1-10-s windows and an AdaBoost ensemble learning algorithm implemented to classify behaviours. Finally, we developed an adjusted count quantification algorithm to predict the prevalence of locomotor play behaviour on a test dataset with low true prevalence (0.27%). Our algorithm identified locomotor play (99.73% accuracy), self-grooming (98.18% accuracy), ruminating (94.47% accuracy), non-nutritive suckling (94.96% accuracy), nutritive suckling (96.44% accuracy), active lying (90.38% accuracy) and non-active lying (90.38% accuracy). Our results detail recommended sampling frequencies, feature selection and window size. The quantification estimates of locomotor play behaviour were highly correlated with the true prevalence (0.97; p < 0.001) with a total overestimation of 18.97%. This study is the first to implement machine learning approaches for multi-class behaviour identification as well as behaviour quantification in calves. This has potential to contribute towards new insights to evaluate the health and welfare in calves by use of wearable sensors.