“…However, a limitation is that Pearson’s correlation does not address the bias of predictions ( González-Recio et al, 2014 ). The RMSE is the most popular measure of prediction error which has been used in several studies on the prediction of cattle BW using image data (e.g., Song et al, 2018 ; Jang et al, 2020 ; Weber et al, 2020 ). RMSE is however scale-dependent and hence, a comparison of results between variables or species is not possible.…”
Section: Discussionmentioning
confidence: 99%
“…Several studies explored the possibility of predicting BW using animal morphological features acquired with novel approaches including computer-vision techniques ( Song et al, 2018 ; Jang et al, 2020 ; Weber et al, 2020 ). Such approaches cover a variety of techniques to generate predictive features based on dairy cows’ morphology.…”
Section: Introductionmentioning
confidence: 99%
“…Such approaches cover a variety of techniques to generate predictive features based on dairy cows’ morphology. In particular, contour data based on 2-dimensional (2D) vision ( Weber et al, 2020 ), thermal vision ( Stajnko et al, 2008 ), stereo vision using multiple calibrated 2D cameras ( Tasdemir et al, 2011 ), and 3-dimensional (3D) vision using one or multiple 3D cameras have been previously explored ( Marinello et al, 2015 ; Salau et al, 2016 ; Song et al, 2018 ; Jang et al, 2020 ). Investigation on the promise of computer vision for the prediction of BW in cattle remains a dynamic research topic where combined efforts promise better prediction accuracy toward mainstreaming image-based systems for prediction and monitoring BW in dairy cattle.…”
Section: Introductionmentioning
confidence: 99%
“…Additionally, previous studies on the prediction of BW in cattle based on digital image data mostly used general linear regression models (e.g., Tasdemir et al, 2011 ; Jang et al, 2020 ). However, the linear regression methods have been shown to have limitations in handling several predictors, as in the case of automated high-throughput phenotyping, with complex and often nonlinear relationships among these predictors ( Comrie, 1997 ).…”
Introduction: The use of automation and sensor-based systems in livestock production allows monitoring of individual cows in real-time and provides the possibility of early warning systems to take necessary management actions against possible anomalies. Among the different RT monitoring parameters, body weight (BW) plays an important role in tracking the productivity and health status.Methods: In this study, various supervised learning techniques representing different families of methods in the machine learning space were implemented and compared for performance in the prediction of body weight from 3D image data in dairy cows. A total of 83,011 records of contour data from 3D images and body weight measurements taken from a total of 914 Danish Holstein and Jersey cows from 3 different herds were used for the predictions. Various metrics including Pearson’s correlation coefficient (r), the root mean squared error (RMSE), and the mean absolute percentage error (MAPE) were used for robust evaluation of the various supervised techniques and to facilitate comparison with other studies. Prediction was undertaken separately within each breed and subsequently in a combined multi-breed dataset.Results and discussion: Despite differences in predictive performance across the different supervised learning techniques and datasets (breeds), our results indicate reasonable prediction accuracies with mean correlation coefficient (r) as high as 0.94 and MAPE and RMSE as low as 4.0 % and 33.0 (kg), respectively. In comparison to the within-breed analyses (Jersey, Holstein), prediction using the combined multi-breed data set resulted in higher predictive performance in terms of high correlation coefficient and low MAPE. Additional tests showed that the improvement in predictive performance is mainly due to increase in data size from combining data rather than the multi-breed nature of the combined data. Of the different supervised learning techniques implemented, the tree-based group of supervised learning techniques (Catboost, AdaBoost, random forest) resulted in the highest prediction performance in all the metrics used to evaluate technique performance. Reported prediction errors in our study (RMSE and MAPE) are one of the lowest in the literature for prediction of BW using image data in dairy cattle, highlighting the promising predictive value of contour data from 3D images for BW in dairy cows under commercial farm conditions.
“…However, a limitation is that Pearson’s correlation does not address the bias of predictions ( González-Recio et al, 2014 ). The RMSE is the most popular measure of prediction error which has been used in several studies on the prediction of cattle BW using image data (e.g., Song et al, 2018 ; Jang et al, 2020 ; Weber et al, 2020 ). RMSE is however scale-dependent and hence, a comparison of results between variables or species is not possible.…”
Section: Discussionmentioning
confidence: 99%
“…Several studies explored the possibility of predicting BW using animal morphological features acquired with novel approaches including computer-vision techniques ( Song et al, 2018 ; Jang et al, 2020 ; Weber et al, 2020 ). Such approaches cover a variety of techniques to generate predictive features based on dairy cows’ morphology.…”
Section: Introductionmentioning
confidence: 99%
“…Such approaches cover a variety of techniques to generate predictive features based on dairy cows’ morphology. In particular, contour data based on 2-dimensional (2D) vision ( Weber et al, 2020 ), thermal vision ( Stajnko et al, 2008 ), stereo vision using multiple calibrated 2D cameras ( Tasdemir et al, 2011 ), and 3-dimensional (3D) vision using one or multiple 3D cameras have been previously explored ( Marinello et al, 2015 ; Salau et al, 2016 ; Song et al, 2018 ; Jang et al, 2020 ). Investigation on the promise of computer vision for the prediction of BW in cattle remains a dynamic research topic where combined efforts promise better prediction accuracy toward mainstreaming image-based systems for prediction and monitoring BW in dairy cattle.…”
Section: Introductionmentioning
confidence: 99%
“…Additionally, previous studies on the prediction of BW in cattle based on digital image data mostly used general linear regression models (e.g., Tasdemir et al, 2011 ; Jang et al, 2020 ). However, the linear regression methods have been shown to have limitations in handling several predictors, as in the case of automated high-throughput phenotyping, with complex and often nonlinear relationships among these predictors ( Comrie, 1997 ).…”
Introduction: The use of automation and sensor-based systems in livestock production allows monitoring of individual cows in real-time and provides the possibility of early warning systems to take necessary management actions against possible anomalies. Among the different RT monitoring parameters, body weight (BW) plays an important role in tracking the productivity and health status.Methods: In this study, various supervised learning techniques representing different families of methods in the machine learning space were implemented and compared for performance in the prediction of body weight from 3D image data in dairy cows. A total of 83,011 records of contour data from 3D images and body weight measurements taken from a total of 914 Danish Holstein and Jersey cows from 3 different herds were used for the predictions. Various metrics including Pearson’s correlation coefficient (r), the root mean squared error (RMSE), and the mean absolute percentage error (MAPE) were used for robust evaluation of the various supervised techniques and to facilitate comparison with other studies. Prediction was undertaken separately within each breed and subsequently in a combined multi-breed dataset.Results and discussion: Despite differences in predictive performance across the different supervised learning techniques and datasets (breeds), our results indicate reasonable prediction accuracies with mean correlation coefficient (r) as high as 0.94 and MAPE and RMSE as low as 4.0 % and 33.0 (kg), respectively. In comparison to the within-breed analyses (Jersey, Holstein), prediction using the combined multi-breed data set resulted in higher predictive performance in terms of high correlation coefficient and low MAPE. Additional tests showed that the improvement in predictive performance is mainly due to increase in data size from combining data rather than the multi-breed nature of the combined data. Of the different supervised learning techniques implemented, the tree-based group of supervised learning techniques (Catboost, AdaBoost, random forest) resulted in the highest prediction performance in all the metrics used to evaluate technique performance. Reported prediction errors in our study (RMSE and MAPE) are one of the lowest in the literature for prediction of BW using image data in dairy cattle, highlighting the promising predictive value of contour data from 3D images for BW in dairy cows under commercial farm conditions.
“…It is important to keep in mind that experiments that used bidimensional sensors and ranged substantial or very good R² (>0.75) combine the dorsal view with the side view or in loco measures (NICHOLAS et al 2018;ZHANG et al 2018;NILCHUEN et al 2021) or using tridimensional sensors as Kinect (GOMES et al 2016;FERNANDES et al 2019;JANG et al 2020) that gives the volume information that is a very for BW prediction. In our experiment the animals were not handled, meaning that had no human contact for any of the data collection, besides our record system used the simplest and cheapest equipment.…”
For the beef cattle system, one of the most valuable information is the body weight that can be linked to animal growth and performance. The bidimensional sensors area is the cheapest technology among all sensors used as a tool to extract information that can be applied to machine learning to predicts value phenotype. This study aimed to predict body weight using video image analysis with simple bidimensional equipment, from the dorsal view of crossbreed beef cattle (1⁄2 Angus x 1⁄2 Nellore) in a finishing system, applying different frame information and machine learning algorithms. The experimental procedures were performed at Federal University of Viçosa. A total of 40 crossbreed steers (1⁄2 Angus x 1⁄2 Nellore) were used, averaging 8 months of age at the beginning of the feedlot trial, and 291.7±23.8 kg and 517.42±54.8kg of initial and final body weight, respectively. The data collection occurred from September (12 Month) to December/2021 (15 month). Body weight (BW) was collected using an automatic Intergado company drink fountain/scale and the video images were collected using cameras Intelbras from the animals’ dorsal view. Three approaches were tried for segmentation of the animals’ dorsal images, however, their color characteristics did not allow do this automatically, so were used ImageJ software to manually do the delimitations, extracting 8 Shape Descriptors. For regression were used 6 machine learning algorithms, Ridge, Least Absolute Shrinkage and Selection Operator (LASSO), Elastic Net (ENET), Multiple Linear Regression (MLR), Adaboost (ADAB) and Random Forest (RF) to construct predictive models, the dataset was split in 70:30 for training and test. The regularizations RIDGE and MLR without AGE as a predictor had similar performance. The AGE addition improved all algorithms, the best metrics results were for ENET and ENET using AGE for a dataset with 5 Frames information (5F) R2=0.68 and 0.76, respectively. Thus, the use of bidimensional sensor in the dorsal view can predict the BW of crossbreed (1⁄2 Angus x 1⁄2 Nellore). Keywords: Correlation. Image Processing Regularization
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.