A lameness scoring system that uses posture and gait to predict dairy cattle reproductive performance
Abstract
Lameness has contributed to reproductive inefficiency and increased the risk of culling in dairy cows. We developed a 5-point lameness scoring system that assessed gait and placed a novel emphasis on back posture. Our objective was to determine if this system predicted future reproductive performance and the risk of culling. The study was conducted at a commercial dairy farm with a history of declining reproductive efficiency and an increasing prevalence of lameness. A total of 66 primipara and pluripara calved, received an initial lameness score, and completed their 60-d voluntary waiting period. The overall prevalence of lameness (mean lameness score >2) was 65.2%. Scoring continued at 4-wk intervals and ceased with conception or culling. The percentage of cows confirmed pregnant and culled was 77.3 and 22.7, respectively.
For each reproductive endpoint, a 2 × 2 table was constructed with lameness score >2 as the positive risk factor and either performance greater than the endpoint mean or being culled as the positive disease or condition. Positive and negative predictive values, relative risk, Chisquare statistic and regression analysis were used to evaluate the data. The positive predictive values for days to first service, days open, breeding herd days, services per pregnancy and being culled were 58, 68, 65, 39 and 35%, respectively. Similarly, the negative predictive values were 79, 96, 100, 96 and 100%, respectively. Except for one reproductive endpoint, the total number of services, all linear regressions were significant at P < 0.01. Having a lameness score >2 predicted that a cow would have extended intervals from calving to first service and to conception, spend or be assigned to (explained herein) more total days in the breeding herd, require more services per pregnancy and be 8.4 times more likely to be culled. We believe that this lameness scoring system effectively identifies lame cows. Observation of the arched-back posture in a standing cow (≥LS 3) should trigger corrective interventions.
References (11)
- LA Lee et al.
Effects of disease on days open assessed by survival analysis
J Dairy Sci
(1989) - JR Pursley et al.
Synchronization of ovulation in dairy cows using PGF2α and GnRH
Theriogenology
(1995) - DJ Sprecher et al.
The educational implications of reproductive problems identified during investigations at Michigan dairy farms
Theriogenology
(1995) - DW Collick et al.
Associations between types of lameness and fertility
Vet Rec
(1989) - S Lucey et al.
The association between lameness and fertility in dairy cows
Vet Rec
(1986)
Cited by (818)
Lameness detection system for dairy cows based on instance segmentation
2024, Expert Systems with ApplicationsLameness is one of the major health problems on dairy farms, which seriously affects dairy cow welfare and increases the risk of premature culling. Accuracy lameness detection ensures timely treatment of hooves and improves the level of health management on dairy farms. Most of the existing lameness detection methods detect one dairy cow, and it is difficult to detect lameness in multiple dairy cows at the same time. In this paper, a lameness detection system based on instance segmentation is constructed to extract the lameness features of multiple dairy cows and automatically detect lameness. First, an improved SOLOv2 network is designed to reduce the semantic gap between low-level and high-level features and improve the precision of dairy cow segmentation. Second, individual matching of dairy cows in different video frames is performed using the Hungarian algorithm. Third, the Canny algorithm is used to extract the neck and back contour features and key gait features of dairy cows. Finally, a deep learning model is constructed, and multiple lameness features are fused to detect the lameness of dairy cows. To evaluate the performance of the constructed system, 172 videos were randomly selected from 246 videos as training videos, and the remaining 74 videos were selected as the test videos. The lameness detection accuracy of the constructed system was 98.65%. The experimental results showed that the constructed system can extract the lameness features of multiple dairy cows at the same time and accurately detect the lameness of dairy cows.
Associations of lameness with milk composition, fatty acid profile, and milk coagulation properties in mid-lactation high-yielding Holstein cows
2024, International Dairy JournalLameness, an animal welfare and profitability related problem in dairy industry, was evaluated on 126 multiparous Estonian Holstein cows on a 5-point scale; score three or above indicated lameness. Milk samples of the cows were analysed for basic composition, content of mineral elements, fatty acid (FA) profile and coagulation properties (MCP). Lame cows had lower milk yield and a lower milk protein and milk fat yield compared to non-lame cows. MCP and content of mineral elements were not affected by lameness. The concentrations of most individual milk FA, their groups formed by chain length, degree of saturation or by origin were not affected by the lameness. Lameness was associated with a higher concentration of odd- and branched-chain fatty acids and C18:0. Milk from lame cows is suitable for consumers as lameness did not impair nutritional value nor MCP of the milk.
A predictive model for hypocalcaemia in dairy cows utilizing behavioural sensor data combined with deep learning
2024, Computers and Electronics in Agriculture(Sub)clinical hypocalcaemia occurs frequently in the dairy industry, and is one of the earliest symptoms of an impaired transition period. Calcium deficiency is accompanied by changes in cows’ daily behavioural variables, which can be measured by sensors. The goal of this study was to construct a predictive model to identify cows at risk of hypocalcaemia in dairy cows using behavioural sensor data. For this study 133 primiparous and 476 multiparous cows from 8 commercial Dutch dairy farms were equipped with neck and leg sensors measuring daily behavioural parameters, including eating, ruminating, standing, lying, and walking behaviour of the 21 days before calving. From each cow, a blood sample was taken within 48 h after calving to measure their blood calcium concentration. Cows with a blood calcium concentration 2.0 mmol/L were defined as hypocalcemic. In order to create a more context based cut-off, a second way of dividing the calcium concentrations into two categories was proposed, using a linear mixed-effects model with a k-Means clustering. Three possible binary predictive models were tested; a logistic regression model, a XgBoost model and a LSTM deep learning model. The models were expanded by adding the following static features as input variables; parity (1, 2 or 3＋), calving season (summer, autumn, winter, spring), day of calcium sampling relative to calving (0, 1 or 2), body condition score and locomotion score. Of the three models, the deep learning model performed best with an area under the receiver operating characteristic curve (AUC) of 0.71 and an average precision of 0.47. This final model was constructed with the addition of the static features, since they improved the model’s tuning AUC with 0.11. The calcium label based on the cut-off categorization method proved to be easier to predict for the models compared to the categorization method with the k-means clustering. This study provides a novel approach for the prediction of hypocalcaemia, and an ameliorated version of the deep learning model proposed in this study could serve as a tool to help monitor herd calcium status and to identify animals at risk for associated transition diseases.
Cow key point detection in indoor housing conditions with a deep learning model
2024, Journal of Dairy ScienceLameness in dairy cattle is a costly and highly prevalent problem that affects all aspects of sustainable dairy production, including animal welfare. Automation of gait assessment would allow monitoring of locomotion in which the cows' walking patterns can be evaluated frequently and with limited labor. With the right interpretation algorithms, this could result in more timely detection of locomotion problems. This in turn would facilitate timely intervention and early treatment, which is crucial to reduce the effect of abnormal behavior and pain on animal welfare. Gait features of dairy cows can potentially be derived from key points that locate crucial anatomical points on a cow's body. The aim of this study is 2-fold: (1) to demonstrate automation of the detection of dairy cows' key points in a practical indoor setting with natural occlusions from gates and races, and (2) to propose the necessary steps to postprocess these key points to make them suitable for subsequent gait feature calculations. Both the automated detection of key points as well as the postprocessing of them are crucial prerequisites for camera-based automated locomotion monitoring in a real farm environment. Side-view video footage of 34 Holstein-Friesian dairy cows, captured when exiting the milking parlor, were used for model development. From these videos, 758 samples of 2 successive frames were extracted. A previously developed deep learning model called T-LEAP was trained to detect 17 key points on cows in our indoor farm environment with natural occlusions. To this end, the dataset of 758 samples was randomly split into a train (n = 22 cows; no. of samples = 388), validation (n = 7 cows; no. of samples = 108), and test dataset (n = 15 cows; no. of samples = 262). The performance of T-LEAP to automatically assign key points in our indoor situation was assessed using the average percentage of correctly detected key points using a threshold of 0.2 of the head length (PCKh0.2). The model's performance on the test set achieved a good result with PCKh0.2: 89% on all 17 key points together. Detecting key points on the back (n = 3 key points) of the cow had the poorest performance PCKh0.2: 59%. In addition to the indoor performance of the model, a more detailed study of the detection performance was conducted to formulate postprocessing steps necessary to use these key points for gait feature calculations and subsequent automated locomotion monitoring. This detailed study included the evaluation of the detection performance in multiple directions. This study revealed that the performance of the key points on a cows' back were the poorest in the horizontal direction. Based on this more in-depth study, we recommend the implementation of the outlined postprocessing techniques to address the following issues: (1) correcting camera distortion, (2) rectifying erroneous key point detection, and (3) establishing the necessary procedures for translating hoof key points into gait features.
A new approach and insights on modelling the impact of production diseases on dairy cow welfare
2024, AnimalAnimal welfare is becoming an important consideration in animal health-related decision-making. Integrating considerations of animal welfare into the decision-making process of farmers involves recognising the significance of health disorder impacts in relation to animal welfare. Yet little research quantifies the impact, making it difficult to include animal welfare in the animal health decision-making process. Quantifying the impact of health disorders on animal welfare is incredibly challenging due to empirical animal-based data collection constraints. An approach to circumvent these constraints is to rely on expert knowledge whereby perceived welfare impairment weights are indicative of the negative welfare effect. In this research, we propose an expertise-based method to quantify the perceived impact of sub-optimal mobility (SOM) on the welfare of dairy cows, because of its welfare importance. We first quantified perceived welfare impairment weights of SOM by eliciting expert knowledge using adaptive conjoint analysis (ACA). Second, using the perceived welfare impairment weights, we derived the perceived welfare disutility (i.e., perceived negative welfare effect) of mobility scores 1–5 (1 = optimal mobility, 5 = severely impaired mobility). Third, using the perceived welfare disutility per mobility score, we quantified the perceived welfare impact at case- and herd-level of SOM for different SOM severity. Results showed that perceived welfare disutility increased with each increase in mobility score. However, the perceived welfare impact of SOM cases with lower mobility scores was higher compared to SOM cases with higher mobility scores. This was because of the longer-lasting duration of the SOM cases with lower mobility scores. Moreover, the perceived herd-level welfare impact was largely due to SOM cases with lower mobility scores because of the longer duration and more frequent incidence compared to the SOM cases with higher mobility scores. These results entail that better welfare of dairy cows with respect to SOM can be achieved if lower mobility scores are detected and treated sooner. Our research demonstrates a novel approach that quantifies the perceived impact of health disorders on animal welfare when empirical evidence is limited.
Quantifying the economic and animal welfare trade-offs of classification models in precision livestock farming for sub-optimal mobility management
2024, Computers and Electronics in AgriculturePrecision livestock farming (PLF) offers a sensor-based management approach to potentially mitigate the negative economic and animal welfare consequences of sub-optimal mobility (SOM). Human-based SOM classification is often done using more than two classes (i.e., mobility scores 1–5, where 1 = optimal and 5 = severely impaired mobility), while binary classification is ultimately used in sensor-based classification. Previous economic research shows that classifying SOM as a binary problem in sensor-based management has little to no economic value while non-binary SOM classification may be more economically beneficial. However, the animal welfare implications of a non-binary SOM classification approach are unknown. In this study, we assess whether economic and welfare gains can be achieved by using 3-class SOM classifiers (i.e., sensors) for sensor-based SOM management compared with the current no-sensor SOM management. With respect to mobility scores 1–5, three SOM classes ( = non-SOM, = SOM, and = severe-SOM) along with two management scenarios, with four different classifiers each, were defined. Mobility scores 1–5 were grouped into one of three SOM classes depending on the classifier. In management scenario one, mobility scores 1 and 2, were grouped to , while mobility score 3 was grouped to and mobility scores 4 and 5 to . In management scenario two, mobility scores 2 and 3 were grouped to . In both management scenarios, alerts for cows classified to SOM class were generated every 7 days based on an alert prioritisation criterion, while alerts for cows classified to SOM class were generated daily. Treatment options followed the generation of either weekly or daily alerts. For each of the eight classifiers (i.e., 4 classifiers per management scenario) 600 classification outcomes were defined. A bio-economic simulation model was used to simulate the economic and welfare effects of the various classifiers and classification outcomes respective of management scenarios. Comparisons were made with a no-classifier scenario. The simulated output data was first analysed using an exploratory approach. Second, a novel method accounting for the highly interactive classification outcomes was developed to quantify the trade-offs in classification outcomes and how they affected the economic and welfare gains. Among the tested classifiers, all showed economic and welfare gains on average. Classifiers with larger separations between non-SOM and SOM classes showed the highest economic gains. Including mobility score 2 into the SOM class showed meaningful welfare gains on average as opposed to when mobility score 2 was included in the non-SOM class . Economic gains were more sensitive to trade-offs in classification outcomes compared to welfare gains. Larger increases in economic gains were often achieved at the cost of smaller reductions in welfare gains for changes in classification outcomes. This study provides valuable insights on designing appropriate 3-class SOM classifiers to be used in practice that could also be beneficial when designing classifiers for health disorders other than SOM. It also demonstrates the value in using simulation models to test classifiers by highlighting interesting classification outcomes that can be further validated in practice.