Across-Species Pose Estimation in Poultry Based on Images Using Deep Learning

Abstract

Animal pose-estimation networks enable automated estimation of key body points in images or videos. This enables animal breeders to collect pose information repeatedly on a large number of animals. However, the success of pose-estimation networks depends in part on the availability of data to learn the representation of key body points. Especially with animals, data collection is not always easy, and data annotation is laborious and time-consuming. The available data is therefore often limited, but data from other species might be useful, either by itself or in combination with the target species. In this study, the across-species performance of animal pose-estimation networks and the performance of an animal pose-estimation network trained on multi-species data (turkeys and broilers) were investigated. Broilers and turkeys were video recorded during a walkway test representative of the situation in practice. Two single-species and one multi-species model were trained by using DeepLabCut and tested on two single-species test sets. Overall, the within-species models outperformed the multi-species model, and the models applied across species, as shown by a lower raw pixel error, normalized pixel error, and higher percentage of keypoints remaining (PKR). The multi-species model had slightly higher errors with a lower PKR than the within-species models but had less than half the number of annotated frames available from each species. Compared to the single-species broiler model, the multi-species model achieved lower errors for the head, left foot, and right knee keypoints, although with a lower PKR. Across species, keypoint predictions resulted in high errors and low to moderate PKRs and are unlikely to be of direct use for pose and gait assessments. A multi-species model may reduce annotation needs without a large impact on performance for pose assessment, however, with the recommendation to only be used if the species are comparable. If a single-species model exists it could be used as a pre-trained model for training a new model, and possibly require a limited amount of new data. Future studies should investigate the accuracy needed for pose and gait assessments and estimate genetic parameters for the new phenotypes before pose-estimation networks can be applied in practice.

Keypoint Detection for Injury Identification during Turkey Husbandry Using Neural Networks

“…Nevertheless, the animals were recorded in the barn in this study and thus in their usual environment. The turkeys were neither stimulated to walk along a corridor [ 43 ] nor placed in a specific photography environment [ 1 ] as in other studies.…”

“…Besides that, the annotation of keypoints on the all-white bodies of the turkeys was already difficult, and thus we could not guarantee that, for instance, the "center of the body" keypoint always had the same position. Doornweerd et al [43] estimated turkey poses based on Figure 6. Combination of KPD generated in this study and injury detection from previous work [16] on the evaluation dataset.…”

“…Finally, they stated that their research provided an appropriate non-invasive method to analyze chicken behavior. More recently, Doornweerd et al [ 43 ] reported the performance of an animal pose estimation network: the network was investigated, trained, and tested on multi-species data from broilers as well as from turkeys. They collected data for pose and gait assessments and evaluated a multi-species model to reduce the required dataset and finally annotation needs.…”

“…They collected data for pose and gait assessments and evaluated a multi-species model to reduce the required dataset and finally annotation needs. Doornweerd et al [ 43 ] recorded the turkeys walking along a corridor, and they used a view from behind the animals, paying particular attention to their locomotion system. They then defined eight key points (head, neck, right and left knee, right and left hock, and right and left foot).…”

128. Automated estimation of pose features in broilers using computer vision

“…Eight keypoints (head, neck, left and right knees, hocks, and feet) were detected using a pre-trained broiler pose estimation deep learning model developed in DeepLabCut (Mathis et al, 2018). Details of the pre-trained broiler pose estimation model are described in detail elsewhere (Doornweerd et al, 2021). However, this pre-trained model struggled with accurately detecting the keypoints of the chickens in the new environment (e.g.…”