Comparative analysis of computer vision algorithms 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 planar aspect of the hoof. DD is associated with massive herd outbreaks of lameness and influences cattle welfare and production. Early detection of DD can lead to prompt treatment and decrease lameness. Computer vision (CV) provides a unique opportunity to improve early detection. The study aims to train and compare applications for the real-time detection of DD in dairy cows. Eight CV models were trained for detection and scoring, compared using performance metrics and inference time, and the best model was automated for real-time detection using images and video. Images were collected from commercial dairy farms while 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 with distinct labels for hyperkeratosis (H) and proliferations (P). Two sets of images were compiled: the first dataset (Dataset 1) containing 1,177 M0/M4H and 1,050 M2/M2P images and the second dataset (Dataset 2) containing 240 M0, 17 M2, 51 M2P, 114 M4H, and 108 M4P images. Models were trained to detect and score DD lesions and compared for precision, recall, and mean average precision (mAP) in addition to inference time in frame per second (FPS). Seven of the nine CV models performed well compared to the ground truth of labeled images using Dataset 1. The six models, Faster R-CNN, Cascade R-CNN, YOLOv3, Tiny YOLOv3, YOLOv4, Tiny YOLOv4, and YOLOv5s achieved an mAP between 0.964 and 0.998, whereas the other two models, SSD and SSD Lite, yielded an mAP of 0.371 and 0.387 respectively. Overall, YOLOv4, Tiny YOLOv4, and YOLOv5s outperformed all other models with almost perfect precision, perfect recall, and a higher mAP. Tiny YOLOv4 outperformed all other models with respect to inference time at 333 FPS, followed by YOLOv5s at 133 FPS and YOLOv4 at 65 FPS. YOLOv4 and Tiny YOLOv4 performed better than YOLOv5s compared to the ground truth using Dataset 2. YOLOv4 and Tiny YOLOv4 yielded a similar mAP of 0.896 and 0.895, respectively. However, Tiny YOLOv4 achieved both higher precision and recall compared to YOLOv4. Finally, Tiny YOLOv4 was able to detect DD lesions on a commercial dairy farm with high performance and speed. The proposed CV tool can be used for early detection and prompt treatment of DD in dairy cows. This result is a step towards applying CV algorithms to veterinary medicine and implementing real-time DD detection on dairy farms.

Sensitivity and specificity of mobility scoring for the detection of foot lesions in pasture-based Irish dairy cows

Lameness is an important production disease in dairy cows worldwide and has detrimental effects on cows' welfare, production, and reproductive performance, thus affecting the sustainability of dairy farming. Timely and effective detection of lameness allows for effective treatment, minimizing progression of disease, and maximizing the prognosis of recovery. Mobility scoring (MSc) is a 4 point (0-3) visual lameness scoring system that is the industry standard in several countries. However, few studies have examined the sensitivity (Se) and specificity (Sp) of MSc to detect foot lesions. The aim of this observational study was to determine the Se and Sp of MSc to detect foot lesions in dairy cattle in a pasture-based system. Five hundred ninety-five primi- and multiparous cows were randomly selected from 12 commercial Irish dairy farms and recruited for the study. Recruited cows were mobility scored and passed through a foot-paring crate where all 4 feet were lifted for examination. The team recorded the anatomical location and severity of any foot lesions present based on appearance only. Then, based on the type and severity of the lesions present, cows were classified according to 3 case definitions case definition 1: Any lesion present; case definition 2: Moderate lesions present (excluding minor lesions expected to have a low probability of affecting gait); and case definition 3: Severe lesions present (only including lesions most likely to result in a detectable gait abnormality). Sensitivity and Sp of MSc was calculated based on a threshold of MSc ≥2, defined as impaired (MSc = 2) or severely impaired (MSc = 3) mobility for each of the 3 case definitions, at the overall level and disaggregated by parity. The overall cow-level lesion prevalence based on the case definition 1 was 0.54 with significant between-herd variation. The overall Se and Sp of MSc for the detection of foot lesions were 0.18 and 0.96, 0.35 and 0.94, 0.43 and 0.94 for the case definitions 1, 2, and 3, respectively. Our findings showed poor Se, but high Sp of MSc for the detection of cows with foot lesions in a pasture-based system.

Efficacy of quaternary ammonium salt-based disinfectant or chelated copper-zinc footbath solution in the treatment of digital dermatitis on one research dairy farm in Taiwan

BACKGROUND

Footbaths can be used to manage digital dermatitis (DD), a common cause of lameness in dairy cattle. Copper sulfate and chelated copper-zinc (CZS) solutions in footbaths are potentially harmful to the environment.

HYPOTHESIS/OBJECTIVES

To determine if a quaternary ammonium salt-based disinfectant (QASD) footbath is as effective as a chelated CZS solution in controlling DD in dairy cows.

ANIMALS

Fifty-one lactating Holstein cows were randomly assigned to one of two treatment groups, with DD status based on the M-stage scoring system and locomotion score balanced between treatment groups.

MATERIALS AND METHODS

The groups were treated with a 1% QASD or a 2.5% chelated CZS. Footbaths were performed once per week for 15 weeks. Logistic regression was used to analyse clinical improvement. All cows received appropriate medical treatment for DD and other hoof diseases.

RESULTS

Clinical improvement rates were 67% in the QASD group and 38% in the CZS group (p = 0.03). Logistic regression analysis showed that the odds (95% confidence interval) for clinical improvement rate in the CZS group were 0.30 (0.095-0.948) times that of the QASD group (p = 0.04). The M0 score in the QASD and CZS groups increased significantly (p < 0.05) at the end of the 15 week study period. In the QASD group, the proportion of M2, M3 and M4 scores were significantly decreased (p < 0.05).

CONCLUSION AND CLINICAL RELEVANCE

Over a 15 week period, QASD for footbathing was associated with a lower prevalence of active DD lesions than when using CZS.

Integrated infrared thermography and accelerometer-based behavior logger as a hoof lesion identification tool in dairy cows with various foot diseases under subtropical climates

Infrared thermography (IRT) can measure a temperature change on the surface of objects, and is widely used as an inflammation or fever detection tool. The objective of this longitudinal study was to investigate the feasibility of detecting hoof lesion cattle using IRT under subtropical climate conditions. The experiment was conducted in two free-stall commercial dairy farms and 502 dairy cows participated between August 2020 and March 2022. Before hoof trimming, the portable IRT was used to measure the maximum temperature of each hoof from three shooting directions, including anterior (hoof coronary band), lateral (hoof lateral coronary band), and posterior (skin between heel and bulbs). In order to evaluate the effect of hoof lesions on the behavior of dairy cows, we also collected behavior data by automated accelerometers. The results indicated that the temperature of hooves with lesions was significantly higher than that of sound hooves in hot environments regardless of the shooting directions (P < 0.0001). In all of three shooting directions, the maximum temperature of feet with severe lesion was significantly higher than those of feet with mild lesion and sound feet (P < 0.05). Cows with lesion feet had lower daily activity and feeding time than sound cows before clinical diagnosis (P < 0.05). Furthermore, we used thresholds of both anterior hoof temperature at 32.05 °C and average daily activity at 410.5 (arbitrary unit/d) as a lame cow detecting tool. The agreement of this integrated tool reached 75% with clinical diagnosis, indicating that this integrated approach may be feasible for practice in dairy farm. In conclusion, IRT has the potential to be used as a hoof lesion detecting tool under subtropical climate conditions when using sound hoof temperature as reference points, and detection precision can be improved when IRT integrated with automated accelerometers as a lame cow detecting tool.

German Farmers' Awareness of Lameness in Their Dairy Herds

Lameness is one of the most challenging problems in the dairy industry. Control is impeded because farmers often underestimate the number of lame cows. The objectives of this study were to assess German farmers' awareness of lameness in their herds and to determine the associations between farmers' awareness and their management practices, farm characteristics as well as with farmers' education, personality traits and attitudes. As a part of a large cross-sectional study, veterinarians visited farms in three structurally different regions of Germany: north (n = 253), east (n = 252), and south (n = 260). The cows (n = 84,998) were scored for locomotion and farmers were asked to estimate the number of cows that were lame or did not walk soundly. The ratio of farmers' estimated prevalence and the veterinarians' observed prevalence (Farmer's Detection Index; FDI) was calculated. The median lameness prevalence assessed by the veterinarians was 23.1, 39.1, and 23.2%, and the median prevalence of lame cows estimated by the farmers was 9.5, 9.5, and 7.1% in the north, east, and south, respectively. On average, farmers were conscious of only 45.3% (north), 24.0% (east), and 30.0% (south) of their lame cows. Farmers managing their herds according to organic principles had a higher FDI than farmers who managed their herds conventionally. Surprisingly, no significant associations between FDI and factors concerning claw health management could be detected. Therefore, increased awareness did not seem to be necessarily linked to improved management. Moreover, the FDI was not significantly associated with farmers' education or herd size. In the south, more extraverted farmers had a lower FDI. Those farmers who totally agreed with the statement, “I am satisfied with my herd's health,” had a lower FDI than farmers who disagreed or were undecided. Moreover, farmers who disagreed or were undecided with the statement, “It affects me to see a cow in pain” had a higher FDI than those farmers who agreed to the statement. The results indicate that poor awareness of lameness was linked to the farmers' attitude and personality. Therefore, new approaches concerning the consultation regarding lameness control, such as the use of Motivational Interviewing, might be useful in the future.

Identifying cow - level factors and farm characteristics associated with locomotion scores in dairy cows using cumulative link mixed models

Lameness is a tremendous problem in intensively managed dairy herds all over the world. It has been associated with considerable adverse effects on animal welfare and economic viability. The majority of studies have evaluated factors associated with gait disturbance by categorising cows into lame and non-lame. This procedure yet entails a loss of information and precision. In the present study, we extend the binomial response to five categories acknowledging the ordered categorical nature of locomotion assessments, which conserves a higher level of information. A cumulative link mixed modelling approach was used to identify factors associated with increasing locomotion scores. The analysis revealed that a low body condition, elevated somatic cell count, more severe hock lesions, increasing parity, absence of pasture access, and poor udder cleanliness were relevant variables associated with higher locomotion scores. Furthermore, distinct differences in the locomotion scores assigned were identified in regard to breed, observer, and season. Using locomotion scores rather than a dichotomised response variable uncovers more refined relationships between gait disturbances and associated factors. This will help to understand the intricate nature of gait disturbances in dairy cows more deeply.

Claw Trimming as a Lameness Management Practice and the Association with Welfare and Production in Dairy Cows

Lameness resulting from claw lesions remains a pressing welfare issue in dairy cows. Claw trimming (CT) is a common practice for prevention and management of clinically lame cows. This review summarizes the results of studies that have investigated various claw trimming (CT) methods, their application in lameness management, and associations with the welfare and production of dairy cows. The papers included in this review fulfilled the following inclusion criteria: published in peer review journal or book chapter within the last 20 years (1999-2019), written in English, and focused on the application of CT for lameness management and the association with either welfare or production variables. Databases used included Google scholar, Web of Science and PubMed. A total of 748 records were assessed and 61 papers were eligible for inclusion and the main objectives and results were used to categorize the results under six topics: CT techniques, association between CT and claw overgrowth/specific claw lesions, timing and frequency of CT, association between CT and behavioral variables, association between CT and physiological parameters, and association between CT and production. The literature findings showed the existence of various CT methods with the common types including the Dutch Five-step, White Line, White Line Atlas, and Kansas techniques. There is data paucity on the efficacy of these techniques in lameness management; however, the slight procedural difference yields varying sole thicknesses and presentations which may influence their prophylactic use. Results regarding the impact of CT on welfare and production were discussed in relation to potential short and long-term benefits. Depending on the lesion type and severity level, CT may induce immediate painful sensation, stress, changes in lying down activities and reduction in milk yield, but the positive impacts were more evident at later stages of lactation following improvement in locomotion score. The majority of the reviewed studies were lacking a detailed description of CT techniques and claw health of the studied animals; thus, reducing the strength of demonstrating CT-related benefits. However, electronic recording of claw health data during every CT visit provides the basis for monitoring hoof health and could assist in curtailing some of these challenges. To elucidate CT-related benefits, certain areas requiring further research were highlighted such as ascertaining the appropriate timing for preventive CT and identifying cows that will benefit more from such intervention during lactation.

Hot topic: Detecting digital dermatitis with computer vision

Digital dermatitis (DD) is linked to severe lameness, infertility, and decreased milk production in cattle. Early detection of DD provides an improved prognosis for treatment and recovery; however, this is extremely challenging on commercial dairy farms. Computer vision (COMV) models can help facilitate early DD detection on commercial dairy farms. The aim of this study was to develop and implement a novel COMV tool to identify DD lesions on a commercial dairy farm. Using a database of more than 3,500 DD lesion images, a model was trained using the YOLOv2 architecture to detect the M-stages of DD. The YOLOv2 COMV model detected DD with an accuracy of 71%, and the agreement was quantified as "moderate" by Cohen's kappa when compared with a human evaluator for the internal validation. In the external validation, the YOLOv2 COMV model detected DD with an accuracy of 88% and agreement was quantified as "fair" by Cohen's kappa. Implementation of COMV tools for DD detection provides an opportunity to identify cows for DD treatment, which has the potential to lower DD prevalence and improve animal welfare on commercial dairy farms.

Bovine digital dermatitis in Victoria, Australia

AIMS

The objectives of this study were to estimate the prevalence of digital dermatitis (DD) in Victoria, Australia, and to investigate which organisms are consistent with typical DD lesions. The prevalence and causative pathogens of DD are not clear yet in Australia and this paper is one of the first to explore these questions in this country.

METHODS

Examination and sampling of limbs was undertaken at three knackeries in Victoria, Australia. Limbs were classified as normal (N), active DD-lesion (A), dried or chronic DD-lesion (D) or suspected case of DD (S). A total of 823 cows were examined. Six skin biopsies were taken at each knackery, from which DNA was extracted for diversity profiling. Histochemical staining of samples was performed on eight of the skin biopsies.

RESULTS

DD was detected in 29.8% of all cows. The prevalence of DD was significantly higher in dairy cows (32.2%) than in beef cows (10.8%). The differential abundance of Treponema-species was significantly increased in dried lesions, compared with the normal skin biopsies. Actinobacteria, Proteobacteria, Firmicutes and Tenericutes were found to be significantly different in abundance in the DD lesions compared with normal skin biopsies. Silver staining of samples showed only mild inflammation and in two samples organisms with morphology consistent with Spirochaetes were detected.

CONCLUSIONS

The calculated prevalence indicates that DD is present in Victoria, Australia. The results of diversity profiling showed that the presence of Treponema-species was significantly different between the samples of DD lesions and normal skin.

Evaluation of arterial digital blood flow using Doppler ultrasonography in healthy dairy cows

BACKGROUND

Local circulatory disturbances have been implicated in the development of foot disorders in cattle. The goals of this study were to evaluate the suitability of the interdigital artery in the pastern region in both hind limbs using pulsed-wave (PW) Doppler ultrasonography and to investigate quantitative arterial blood flow variables at that site in dairy cows. An Esaote MyLabOne ultrasound machine with a 10-MHz linear transducer was used to assess blood flow in the interdigital artery in the pastern region in both hind limbs of 22 healthy German Holstein cows. The cows originated from three commercial farms and were restrained in a standing hoof trimming chute without sedation.

RESULTS

A PW Doppler signal suitable for analysis was obtained in 17 of 22 cows. The blood flow profiles were categorised into four curve types, and the following quantitative variables were measured in three uniform cardiac cycles: vessel diameter, pulse rate, maximum systolic velocity, maximum diastolic velocity, end-diastolic velocity, reverse velocity, maximum time-averaged mean velocity, blood flow rate, resistance index and persistence index. The measurements did not differ among cows from the three farms. Maximum systolic velocity, vessel diameter and pulse rate did not differ but other variables differed significantly among blood flow profiles.

CONCLUSIONS

Differences in weight-bearing are thought to be responsible for the normal variability of blood flow profiles in healthy cows. The scanning technique used in this report for evaluation of blood flow in the interdigital artery appears suitable for further investigations in healthy and in lame cows.

Evaluation of arterial digital blood flow using Doppler ultrasonography in healthy dairy cows

BACKGROUND

Local circulatory disturbances have been implicated in the development of foot disorders in cattle. The goals of this study were to evaluate the suitability of the interdigital artery in the pastern region in both hind limbs using pulsed-wave (PW) Doppler ultrasonography and to investigate quantitative arterial blood flow variables at that site in dairy cows. An Esaote MyLabOne ultrasound machine with a 10-MHz linear transducer was used to assess blood flow in the interdigital artery in the pastern region in both hind limbs of 22 healthy German Holstein cows. The cows originated from three commercial farms and were restrained in a standing hoof trimming chute without sedation.

RESULTS

A PW Doppler signal suitable for analysis was obtained in 17 of 22 cows. The blood flow profiles were categorised into four curve types, and the following quantitative variables were measured in three uniform cardiac cycles: vessel diameter, pulse rate, maximum systolic velocity, maximum diastolic velocity, end-diastolic velocity, reverse velocity, maximum time-averaged mean velocity, blood flow rate, resistance index and persistence index. The measurements did not differ among cows from the three farms. Maximum systolic velocity, vessel diameter and pulse rate did not differ but other variables differed significantly among blood flow profiles.

CONCLUSIONS

Differences in weight-bearing are thought to be responsible for the normal variability of blood flow profiles in healthy cows. The scanning technique used in this report for evaluation of blood flow in the interdigital artery appears suitable for further investigations in healthy and in lame cows.

Use of Extended Characteristics of Locomotion and Feeding Behavior for Automated Identification of Lame Dairy Cows

This study was carried out to detect differences in locomotion and feeding behavior in lame (group L; n = 41; gait score ≥ 2.5) and non-lame (group C; n = 12; gait score ≤ 2) multiparous Holstein cows in a cross-sectional study design. A model for automatic lameness detection was created, using data from accelerometers attached to the hind limbs and noseband sensors attached to the head. Each cow’s gait was videotaped and scored on a 5-point scale before and after a period of 3 consecutive days of behavioral data recording. The mean value of 3 independent experienced observers was taken as a definite gait score and considered to be the gold standard. For statistical analysis, data from the noseband sensor and one of two accelerometers per cow (randomly selected) of 2 out of 3 randomly selected days was used. For comparison between group L and group C, the T-test, the Aspin-Welch Test and the Wilcoxon Test were used. The sensitivity and specificity for lameness detection was determined with logistic regression and ROC-analysis. Group L compared to group C had significantly lower eating and ruminating time, fewer eating chews, ruminating chews and ruminating boluses, longer lying time and lying bout duration, lower standing time, fewer standing and walking bouts, fewer, slower and shorter strides and a lower walking speed. The model considering the number of standing bouts and walking speed was the best predictor of cows being lame with a sensitivity of 90.2% and specificity of 91.7%. Sensitivity and specificity of the lameness detection model were considered to be very high, even without the use of halter data. It was concluded that under the conditions of the study farm, accelerometer data were suitable for accurately distinguishing between lame and non-lame dairy cows, even in cases of slight lameness with a gait score of 2.5.