Location: Poultry Research
Title: Automated measurement of broiler stretching behaviors under four stocking densities via faster region-based convolutional neural networkAuthor
LI, GUOMING - Mississippi State University | |
ZHAO, YANG - Mississippi State University | |
PORTER, ZACH - Mississippi State University | |
Purswell, Joseph - Jody |
Submitted to: Animal-The International Journal of Animal Biosciences
Publication Type: Peer Reviewed Journal Publication Acceptance Date: 8/17/2020 Publication Date: 1/10/2021 Citation: Li, G., Zhao, Y., Porter, Z., Purswell, J.L. 2021. Automated measurement of broiler stretching behaviors under four stocking densities via faster region-based convolutional neural network. Animal-The International Journal of Animal Biosciences. 15(1):1751-7311. https://doi.org/10.1016/j.animal.2020.100059. DOI: https://doi.org/10.1016/j.animal.2020.100059 Interpretive Summary: Automatic monitoring of animal behavior could allow for continuous assessment of animal welfare and provide a means to integrate welfare and comfort into environmental control and flock management systems. Stretching behavior is a comfort behavior in broilers that could be used in this capacity, but there is no means to automatically monitor stretching behavior under representative production practices. A neural network based behavior detection algorithm was developed to identify and quantify stretching behavior of broilers under multiple stocking densities ranging from 27 to 39 kg/m2¬. The algorithm was able to reliably detect stretching behavior for all stocking densities and multiple bird age. Broilers exhibited stretching behavior from 230-533 sec/day and exhibited less stretching during the first and last two hours of the photoperiod. Results of this study show that the stretching behavior detection algorithm is a useful tool for broiler stretching detection. Technical Abstract: Stretching behavior is one of broiler comfort behaviors that could be used for animal welfare assessment. However, there is currently no methodology for automatic monitoring of stretching behavior under representative production practices. The objectives of this study were to (1) develop a faster region-based convolutional neural network (faster R-CNN) stretching behavior detector for broiler stretching behavior detection, (2) evaluate broiler stretching behaviors under stocking densities (SDs) of 27 (27SD), 29 (29SD), 33 (33SD), and 39 kg·m-2 (39SD) and at weeks 4 and 5 of bird age, and (3) examine the temporal and spatial distribution of broiler stretching behaviors. The results show that the precision, recall, specificity, and accuracy were over 86% on broiler stretching detection across all SDs and bird ages using the faster R-CNN stretching behavior detector. Broilers spent 230-533 sec stretching every day and showed more stretching behaviors under the 29SD, 33SD, and 39SD in week 4 and under the 29SD and 33SD in week 5, as compared to other SDs. They performed less stretching in a couple of hours after light ON and before light OFF but preferred to stretch in areas with less traffic and disturbance, i.e. along the fences and away from the inspection aisle. It is concluded that the stretching behavior detector had acceptable performance in detecting broiler stretching, thus being a useful tool for broiler stretching detection. Broiler stretching behavior is affected by stocking density and bird age and shows temporal and spatial variations. |