Benchmarking analysis of computer vision algorithms on edge devices for the real-time detection of digital dermatitis in dairy cows.

Digital dermatitis (DD) is a bovine claw disease responsible for ulcerative lesions on the coronary band of the foot. It causes significant animal welfare and economic losses to the cattle industry. Early detection of DD can lead to prompt treatment and decrease lameness. Current detection and staging methods require a trained individual to evaluate the interdigital space on each foot for clinical signs of DD. Computer vision (CV), a type of artificial intelligence for image analysis, has demonstrated promising results on object detection tasks. However, farms require robust solutions that can be deployed in harsh conditions including dust, debris, humidity, precipitation, other equipment issues. The study aims to train, deploy, and benchmark DD detection models on edge devices. Images were collected from commercial dairy farms with the camera facing the interdigital space on the plantar surface of the foot. Images were scored for M-stages of DD by a trained investigator using the M-stage DD classification system. Models were trained to detect and score DD lesions and embedded on an edge device. The Tiny YOLOv4 model deployed on a CV specific integrated camera module connected to a single board computer achieved a mean average precision (mAP) of 0.895, an overall prediction accuracy of 0.873, and a Cohen's kappa of 0.830 for agreement between the computer vision model and the trained investigator. The model reached a final inference speed of 40 frames per second (FPS) and ran stably without any interruptions. The CV model was able to detect DD lesions on an edge device with high performance and speed. The CV tool can be used for early detection and prompt treatment of DD in dairy cows. Real-time detection of DD on edge device will improve health outcomes, while simultaneously decreasing labor costs. We demonstrate that the deployed model can be a low-power and portable solution for real-time detection of DD on dairy farms. This result is a step towards applying CV algorithms to veterinary medicine and implementing real-time detection of health outcomes in precision farming.

DairyCoPilot-Automated data compilation and analysis tools for DairyComp data assets.

Modern dairy farm management requires meaningful data and careful analysis to maximize profitability, cow health, and welfare. Current data platforms, such as DairyComp, lack robust integrated data analysis tools. Producers and consultants need dedicated tools to turn collected data sets into assets for informed decision-making processes. The DairyCoPilot app allows users to rapidly extract health and production data from DairyComp, then compile and analyze the data using a menu-driven point-and-click approach. Prospects for training consultants in applied data analysis skills make DairyCoPilot a tool to identify farm management bottlenecks with less time spent for data analysis, improving cow health, and dairy production. The DairyCoPilot Dashboard R Shiny application is published using RStudio Connect: https://connect.doit.wisc.edu/dairy-copilot/.

Computer vision model for the detection of canine pododermatitis and neoplasia of the paw.

BACKGROUND: Artificial intelligence (AI) has been used successfully in human dermatology. AI utilises convolutional neural networks (CNN) to accomplish tasks such as image classification, object detection and segmentation, facilitating early diagnosis. Computer vision (CV), a field of AI, has shown great results in detecting signs of human skin diseases. Canine paw skin diseases are a common problem in general veterinary practice, and computer vision tools could facilitate the detection and monitoring of disease processes. Currently, no such tool is available in veterinary dermatology. ANIMALS: Digital images of paws from healthy dogs and paws with pododermatitis or neoplasia were used. OBJECTIVES: We tested the novel object detection model Pawgnosis, a Tiny YOLOv4 image analysis model deployed on a microcomputer with a camera for the rapid detection of canine pododermatitis and neoplasia. MATERIALS AND METHODS: The prediction performance metrics used to evaluate the models included mean average precision (mAP), precision, recall, average precision (AP) for accuracy and frames per second (FPS) for speed. RESULTS: A large dataset labelled by a single individual (Dataset A) used to train a Tiny YOLOv4 model provided the best results with a mean mAP of 0.95, precision of 0.86, recall of 0.93 and 20 FPS. CONCLUSIONS AND CLINICAL RELEVANCE: This novel object detection model has the potential for application in the field of veterinary dermatology.

[Use of milk haptoglobin concentration as an indicator in animal health monitoring of dairy cows]. / Nutzung der Haptoglobinkonzentration in Milch als Indikator für das Tiergesundheitsmonitoring bei Milchkühen.

OBJECTIVE: The aim of the present study was to investigate relationships between elevated haptoglobin concentrations in milk and clinical as well as laboratory parameters in early lactating dairy cows. Furthermore, cut-off values should be identified for the differentiation of healthy and affected animals. MATERIAL AND METHODS: 1462 dairy cows between 5.-65. days in milk were examined on 68 Bavarian farms. Milk and blood samples were taken once a week for a 7-week period per farm and body-condition-scoring, backfat thickness measurement and Metricheck examination, to evaluate uterine health, were performed. Milk samples were analysed for milk fat, milk protein, lactose, urea, ß-hydroxybutyrate and non-esterified fatty acids (indirect measurement, based on IR spectra), cell count, and milk haptoglobin. Blood samples were analysed for creatinine, aspartate aminotransferase, gamma-glutamyl transferase, glutamate dehydrogenase, total protein, albumin, creatine kinase, ß-hydroxybutyrate, non-esterified fatty acids, and blood haptoglobin.Cluster analyses were performed to determine cut-off values for haptoglobin. RESULTS: Besides milk haptoglobin (µg/ml) and blood haptoglobin (µg/ml), cell count (cells/ml milk), milk fat (%), milk protein (%), non-esterified fatty acids in blood (mmol/l), lactation number, days in milk, breed, season, and milk yield (kg) were included as significant input variables (p<0.005) in the cluster analyses. Cluster analysis, using k-means resp. k-prototypes algorithms, resulted in 5 (clusters 1-5 M1) resp. 4 different clusters (clusters 0-3 M2 and 0-3 B).A cut-off value of 0.5 µg/ml haptoglobin in milk was determined for the differentiation of healthy and affected animals. CONCLUSION AND CLINICAL RELEVANCE: As milk is an easily available substrate, routine determination of haptoglobin in milk might be a suitable parameter for animal health monitoring. Using the detected cut-off value, apparently healthy animals with subclinical inflammatory diseases can be identified more quickly.

Differentiating between metabolic health statuses in Simmental cows and describing related milk fatty acids and relevant associated factors.

The aim of this observational study was to examine differences in milk fatty acid (FA) concentrations for different metabolic health statuses and for associated factors-specifically to examine with which FA concentrations an increased risk for developing a poor metabolic adaptation syndrome (PMAS) was associated. During weekly visits over 51 wk, blood samples were collected from cows between 5 and 50 days in milk. The farmer collected corresponding milk samples from all voluntary milkings. The analysis was performed on n = 2,432 samples from n = 553 Simmental cows. The observations were assigned to five different cow types (healthy, clever, athletic, hyperketonemic, and PMAS, representing five metabolic health statuses), based on the thresholds of 0.7 mmol/L, 1.2 mmol/L, and 1.4 for the concentrations of ß-hydroxybutyrate and nonesterified fatty acids and for the milk fat-to-protein ratio, respectively. Linear regression models using the predictor variables cow type, parity, week of lactation, and milk yield as fixed effects were developed using a stepwise forward selection to test for significant associations of predictor variables regarding FA concentrations in milk. There was a significant interaction term found between PMAS cows and parity compared to healthy cows for C18:1 (P < 0.001) and for C18:0 (P < 0.01). It revealed higher concentrations for PMAS in primiparous and multiparous cows compared to healthy cows, the slope being steeper for primiparous cows. Further, an interaction term was found between PMAS cows and milk yield compared to healthy cows and milk yield for C16:0 (P < 0.05), revealing a steeper slope for the decrease of C16:0 concentrations with increasing milk yield for PMAS compared to healthy cows. The significant associations and interaction terms between cow type, parity, week of lactation, and milk yield as predictor variables and C16:0, C18:0, and C18:1 concentrations suggest excellent opportunities for cow herd health screening during the early postpartum period.

Effects of GnRH and hCG administration during early luteal phase on estrous cycle length, expression of estrus and fertility in lactating dairy cows.

Our objective was to determine the effects of a single treatment of human chorionic gonadotropin (hCG) or GnRH from d 5 to 7 of the estrous cycle on cycle length, expression of estrus and fertility in lactating dairy cows. Lactating Holstein cows (n = 354) located in Farm 1 and lactating Jersey cows located in Farm 2 (n = 210) detected in estrus by an Automated Activity Monitor (AAM) system from 27 to 50 days in milk (DIM) were randomly assigned to receive one of three treatments from d 5 to 7 of the estrous cycle: control (untreated; CON; Holstein, n = 111; Jersey, n = 66), GnRH (86 µg gonadorelin acetate im; Holstein, n = 116; Jersey, n = 75), or hCG (3,300 IU im; Holstein, n = 127; Jersey, n = 69). Ovaries were scanned with ultrasound in a random subgroup of cows (Holstein/Farm 1, n = 147; Jersey/Farm 2, n = 94) on the day of treatment and 3 or 4 d later to determine ovulation. Estrus was detected after treatment by an AAM, and peak activity and heat index were recorded. A random subgroup of cows observed in estrus after treatment received first AI from 51 to 80 DIM (Holstein, n = 208; Jersey, n = 138). Pregnancy diagnoses were performed by transrectal ultrasonography at 37 ± 3 d post-AI. Holstein and Jersey cows treated with GnRH and hCG had an increased (P < 0.05) ovulatory response compared with controls. Human chorionic gonadotropin decreased (74%; P = 0.05) and GnRH tended to reduce (75%; P = 0.07) the proportion of multiparous Holstein cows returning to estrus compared with CON (86%). Cows treated with hCG had a longer (P < 0.01) estrous cycle length (24.6 ± 0.3 d, Holstein; 23.0 ± 0.3 d, Jersey) compared with CON cows (22.7 ± 0.3 d, Holstein; 21.3 ± 0.3 d Jersey) and GnRH (22.9 ± 0.3 d, Holstein; 21.1 ± 0.3 d Jersey). The percentage of cows with high (≥80) peak activity and heat index did not differ (P > 0.50) between treatments, and milk production did not affect (P > 0.65) the duration of estrus. Pregnancy per AI (P/AI) was not affected by treatments in Holstein (P = 0.93; CON: 34.3%, GnRH: 35.4%, and hCG: 31.5%) and in multiparous Jersey cows (P = 0.35; CON: 34.3%, GnRH: 35.4%, and hCG: 31.5%), but hCG had greater (P = 0.03; 55%) P/AI than GnRH (30.0%) and a trend (P = 0.06) for greater P/AI than CON (33.3%) in primiparous Jersey cows. In summary, inducing the formation of an accessory corpus luteum from d 5 to 7 of the estrous cycle with hCG reduced expression of estrus in multiparous Holstein cows. Moreover, hCG increased estrous cycle length in Holstein and Jersey cows, and it did not affect first service P/AI at 37 ± 3 d post-AI in Holstein and multiparous Jersey lactating cows. However, hCG increased P/AI in primiparous Jersey cows. Future research with a larger number of cows is needed to confirm these intriguing fertility results.