Development and validation of a visual image analysis for monitoring the body size of sheep

Machine Vision Analysis of Ujumqin Sheep's Walking Posture and Body Size

The ability to recognize the body sizes of sheep is significantly influenced by posture, especially without artificial fixation, leading to more noticeable changes. This study presents a recognition model using the Mask R-CNN convolutional neural network to identify the sides and backs of sheep. The proposed approach includes an algorithm for extracting key frames through mask calculation and specific algorithms for head-down, head-up, and jumping postures of Ujumqin sheep. The study reported an accuracy of 94.70% in posture classification. We measured the body size parameters of Ujumqin sheep of different sexes and in different walking states, including observations of head-down and head-up. The errors for the head-down position of rams, in terms of body slanting length, withers height, hip height, and chest depth, were recorded as 0.08 ± 0.06, 0.09 ± 0.07, 0.07 ± 0.05, and 0.12 ± 0.09, respectively. For rams in the head-up position, the corresponding errors were 0.06 ± 0.05, 0.06 ± 0.05, 0.07 ± 0.05, and 0.13 ± 0.07, respectively. The errors for the head-down position of ewes, in terms of body slanting length, withers height, hip height, and chest depth, were recorded as 0.06 ± 0.05, 0.09 ± 0.08, 0.07 ± 0.06, and 0.13 ± 0.10, respectively. For ewes in the head-up position, the corresponding errors were 0.06 ± 0.05, 0.08 ± 0.06, 0.06 ± 0.04, and 0.16 ± 0.12, respectively. The study observed that sheep walking through a passage exhibited a more curved knee posture compared to normal measurements, often with a lowered head. This research presents a cost-effective data collection scheme for studying multiple postures in animal husbandry.

Estimation of the genetic parameters of sheep growth traits based on machine vision acquisition

In the realm of animal phenotyping, manual measurements are frequently utilised. While machine-generated data show potential for enhancing high-throughput breeding, additional research and validation are imperative before incorporating them into genetic evaluation processes. This research presents a method for managing meat sheep and collecting data, utilising the Sheep Data Recorder system for data input and the Sheep Body Size Collector system for image capture. The study aimed to investigate the genetic parameter changes of growth traits in Ujumqin sheep by comparing machine-generated measurements with manual measurements. The dataset consisted of 552 data points from the offspring of 75 breeding rams and 399 breeding ewes. Six distinct random regression models were assessed to pinpoint the most suitable model for estimating genetic parameters linked to growth traits. These models were distinguished based on the inclusion or exclusion of maternal genetic effects, maternal permanent environmental effects, and covariance between maternal and direct genetic effects. Fixed factors such as individual age, individual sex, and ewe age were taken into account in the analysis. The genetic parameters for the yearling growth traits of Ujumqin sheep were calculated using ASReml software. The Akaike information criterion, the Bayesian information criterion, and fivefold cross-validation were employed to identify the optimal model. Research findings indicate that the most accurate models for manually measured data revealed heritability estimates of 0.12 ± 0.15 for BW, 0.05 ± 0.07 for body slanting length, 0.03 ± 0.07 for withers height, 0.15 ± 0.12 for hip height, 0.11 ± 0.11 for chest depth, 0.13 ± 0.13 for shoulder width, and 0.53 ± 0.15 for chest circumference. The optimal models for machine-predicted data showed heritability estimates of 0.1 ± 0.09 for body slanting length, 0.14 ± 0.12 for withers height, 0.55 ± 0.15 for hip height, 0.34 ± 0.15 for chest depth, 0.26 ± 0.15 for shoulder width, and 0.47 ± 0.16 for chest circumference. In manually measured data, genetic correlations ranged from 0.35 to 0.99, while phenotypic correlations ranged from 0.07 to 0.90. In machine data, genetic correlations ranged from -0.05 to 0.99, while phenotypic correlations ranged from 0.03 to 0.84. The results suggest that machine-based estimations may lead to an overestimation of heritability, but this discrepancy does not impact the selection of breeding models.

Body Composition Estimation in Breeding Ewes Using Live Weight and Body Parameters Utilizing Image Analysis

Farmers are continually looking for new, reliable, objective, and non-invasive methods for evaluating the conditions of ewes. Live weight (LW) and body condition score (BCS) are used by farmers as a basis to determine the condition of the animal. Body composition is an important aspect of monitoring animal condition. The body composition is the amount of fat, muscle, and bone; knowing the amount of each is important because the information can be used for better strategic management interventions. Experiments were conducted to establish the relationship between body composition and body parameters at key life stages (weaning and pre-mating), using measurements automatically determined by an image processing application for 88 Coopworth ewes. Computerized tomography technology was used to determine the body composition. Multivariate linear regression (MLR), artificial neural network (ANN), and regression tree (RT) statistical analysis methods were used to develop a relationship between the body parameters and the body composition. A subset of data was used to validate the predicted model. The results showed a correlation between fat, muscle, and bone determined by CT and the fat, muscle, and bone weight estimated by the live weight and body parameters calculated using the image processing application, with r² values of 0.90 for fat, 0.72 for muscle, and 0.50 for bone using ANN. From these results, farmers can utilize these measurements to enhance nutritional and management practices.

Identification of Candidate Genes Associated with Yak Body Size Using a Genome-Wide Association Study and Multiple Populations of Information

Yaks have evolved several breeds or genetic resources owing to their geographical and ecological environment, and investigating the genetic construction of body size among breeds is key for breeding. Here, a genome-wide association study (GWAS) was performed for five body size traits in 31 yak breeds and genetic resources. The information from clustering individuals according to their habitats was used for kinship grouping in the compressed mixed linear model (CMLM). We named this approach the pCMLM method. A total of 3,584,464 high-quality single nucleotide polymorphisms (SNPs) were obtained, and six markers were found to be significantly associated with height by pCMLM. Four candidate genes, including FXYD6, SOHLH2, ADGRB2, and OSBPL6, were identified. Our results show that when CMLM cannot identify optimal clustering groups, pCMLM can provide sufficient associated results based on population information. Moreover, this study provides basic information on the gene localization of quantitative traits of body size among yak breeds.

Automatic measurement and prediction of Chinese Grown Pigs weight using multilayer perceptron neural networks

The knowledge of body size/weight is necessary for the general growth enhancement of swine as well as for making informed decisions that concern their health, productivity, and yield. Therefore, this work aims to automate the collection of pigs' body parameters using images from Kinect V2 cameras, and the development of Multilayer Perceptron Neural Network (MLP NN) models to predict their weight. The dataset obtained using 3D light depth cameras contains 9980 pigs across the S21 and S23 breeds, and then grouped into 70:15:15 training, testing, and validation sets, respectively. Initially, two MLP models were built and evaluations revealed that model 1 outperformed model 2 in predicting pig weights, with root mean squared error (RMSE) values of 5.5 and 6.0 respectively. Moreover, employing a normalized dataset, two new models (3 and 4) were developed and trained. Subsequently, models 2, 3, and 4 performed significantly better with a RMSE value of 5.29 compared to model 1, which has a RMSE value of 6.95. Model 3 produced an intriguing discovery i.e. accurate forecasting of pig weights using just two characteristics, age and abdominal circumference, and other error values show corresponding results.

Accuracy of image analysis for linear zoometric measurements in dromedary camels

The present study was designed to verify the effectiveness of the image analysis method for body measurement in dromedary camel compared to manual measurements as a reference method. To achieve this aim, twenty-one linear body measurements were estimated on 59 adult Sahraoui dromedary camels (22 males and 37 females) with a normal clinical condition by using a measuring stick or vernier caliper (standard method). On the other hand, image analysis on profile, front, or behind photographs was processed using Axiovision Software. Overall mean comparison, relative error, variance, Pearson's correlation coefficient, and coefficient of variance showed that the image analysis method was accurate in relation to the manual measurement. Furthermore, image analysis results indicated relevant accuracy (bias correction factor, C_b ≈1) and precision (Pearson ρ ≈1) which were significantly correlated with the results of the reference method (Lin's concordance correlation coefficients r_ccc ≈ 1). According to Bland-Altman upper and lower limits of agreement, the concordance was estimated between 93.22 and 98.3%. Passing-Bablok regression showed a good relationship between the results of the two methods displaying no significant systematic and proportional bias. The image analysis method for linear body measurements in dromedary camel showed results that are in agreement with the manual measuring method. Therefore, the image analysis could be considered a valid tool for camel conformation trait studies.

ASAS-NANP SYMPOSIUM: Applications of machine learning for livestock body weight prediction from digital images

Monitoring, recording, and predicting livestock body weight (BW) allows for timely intervention in diets and health, greater efficiency in genetic selection, and identification of optimal times to market animals because animals that have already reached the point of slaughter represent a burden for the feedlot. There are currently two main approaches (direct and indirect) to measure the BW in livestock. Direct approaches include partial-weight or full-weight industrial scales placed in designated locations on large farms that measure passively or dynamically the weight of livestock. While these devices are very accurate, their acquisition, intended purpose and operation size, repeated calibration and maintenance costs associated with their placement in high-temperature variability, and corrosive environments are significant and beyond the affordability and sustainability limits of small and medium size farms and even of commercial operators. As a more affordable alternative to direct weighing approaches, indirect approaches have been developed based on observed or inferred relationships between biometric and morphometric measurements of livestock and their BW. Initial indirect approaches involved manual measurements of animals using measuring tapes and tubes and the use of regression equations able to correlate such measurements with BW. While such approaches have good BW prediction accuracies, they are time consuming, require trained and skilled farm laborers, and can be stressful for both animals and handlers especially when repeated daily. With the concomitant advancement of contactless electro-optical sensors (e.g., 2D, 3D, infrared cameras), computer vision (CV) technologies, and artificial intelligence fields such as machine learning (ML) and deep learning (DL), 2D and 3D images have started to be used as biometric and morphometric proxies for BW estimations. This manuscript provides a review of CV-based and ML/DL-based BW prediction methods and discusses their strengths, weaknesses, and industry applicability potential.

Measuring body dimensions of leopards (Panthera pardus) from camera trap photographs

Measurement of body dimensions of carnivores usually requires the chemical immobilization of subjects. This process can be dangerous, costly and potentially harmful to the target individuals. Development of an alternative, inexpensive, and non-invasive method therefore warrants attention. The objective of this study was to test whether it is possible to obtain accurate measurements of body dimensions of leopards from camera trap photographs. A total of 10 leopards (Panthera pardus) were captured and collared at Malilangwe Wildlife Reserve, Zimbabwe from May 7 to June 20, 2017 and four body measurements namely shoulder height, head-to-tail, body, and tail length were recorded. The same measurements were taken from 101 scaled photographs of the leopards recorded during a baited-camera trapping (BCT) survey conducted from July 1 to October 22, 2017 and differences from the actual measurements calculated. Generalized Linear Mixed Effects Models were used to determine the effect of type of body measurement, photographic scale, posture, and sex on the accuracy of the photograph-based measurements. Type of body measurement and posture had a significant influence on accuracy. Least squares means of absolute differences between actual and photographic measurements showed that body length in the level back-straight forelimb-parallel tail posture was measured most accurately from photographs (2.0 cm, 95% CI [1.5–2.7 cm]), while head-to-tail dimensions in the arched back-bent forelimb-parallel tail posture were least accurate (8.3 cm, 95% CI [6.1–11.2 cm]). Using the BCT design, we conclude that it is possible to collect accurate morphometric data of leopards from camera trap photographs. Repeat measurements over time can provide researchers with vital body size and growth rate information which may help improve the monitoring and management of species of conservation concern, such as leopards.

Non-Contact Body Measurement for Qinchuan Cattle with LiDAR Sensor

The body dimension measurement of large animals plays a significant role in quality improvement and genetic breeding, and the non-contact measurements by computer vision-based remote sensing could represent great progress in the case of dangerous stress responses and time-costing manual measurements. This paper presents a novel approach for three-dimensional digital modeling of live adult Qinchuan cattle for body size measurement. On the basis of capturing the original point data series of live cattle by a Light Detection and Ranging (LiDAR) sensor, the conditional, statistical outliers and voxel grid filtering methods are fused to cancel the background and outliers. After the segmentation of K-means clustering extraction and the RANdom SAmple Consensus (RANSAC) algorithm, the Fast Point Feature Histogram (FPFH) is put forward to get the cattle data automatically. The cattle surface is reconstructed to get the 3D cattle model using fast Iterative Closest Point (ICP) matching with Bi-directional Random K-D Trees and a Greedy Projection Triangulation (GPT) reconstruction method by which the feature points of cattle silhouettes could be clicked and calculated. Finally, the five body parameters (withers height, chest depth, back height, body length, and waist height) are measured in the field and verified within an accuracy of 2 mm and an error close to 2%. The experimental results show that this approach could be considered as a new feasible method towards the non-contact body measurement for large physique livestock.