Developing and Evaluating Poultry Preening Behavior Detectors via Mask Region-Based Convolutional Neural Network

SIMPLE SUMMARY: Preening is poultry grooming and comfort behavior to keep plumages in good conditions. Automated tools to continuously monitor poultry preening behaviors remain to be developed. We developed and evaluated hen preening behavior detectors using a mask region-based convolutional neural network (mask R-CNN). Thirty Hy-line brown hens kept in an experimental pen were used for the detector development. Different backbone architectures and hyperparameters (e.g., pre-trained weights, image resizers, etc.) were evaluated to determine the optimal ones for detecting hen preening behaviors. A total of 1700 images containing 12,014 preening hens were used for model training, validation and testing. Our results show that the final performance of detecting hen preening was over 80% for precision, recall, specificity, accuracy, F1 score and average precision, indicating decent detection performance. The mean intersection over union (MIOU) was 83.6–88.7%, which shows great potential for segmenting objects of concern. The detectors with different architectures and hyperparameters performed differently for detecting preening birds and thus we need to carefully adjust these parameters to obtain a robust deep learning detector. In summary, deep learning techniques may have a great ability to automatically monitor poultry behaviors and assist welfare-oriented poultry management. ABSTRACT: There is a lack of precision tools for automated poultry preening monitoring. The objective of this study was to develop poultry preening behavior detectors using mask R-CNN. Thirty 38-week brown hens were kept in an experimental pen. A surveillance system was installed above the pen to record images for developing the behavior detectors. The results show that the mask R-CNN had 87.2 ± 1.0% MIOU, 85.1 ± 2.8% precision, 88.1 ± 3.1% recall, 95.8 ± 1.0% specificity, 94.2 ± 0.6% accuracy, 86.5 ± 1.3% F1 score, 84.3 ± 2.8% average precision and 380.1 ± 13.6 ms·image(−1) processing speed. The six ResNets (ResNet18-ResNet1000) had disadvantages and advantages in different aspects of detection performance. Training parts of the complex network and transferring some pre-trained weights from the detectors pre-trained in other datasets can save training time but did not compromise detection performance and various datasets can result in different transfer learning efficiencies. Resizing and padding input images to different sizes did not affect detection performance of the detectors. The detectors performed similarly within 100–500 region proposals. Temporal and spatial preening behaviors of individual hens were characterized using the trained detector. In sum, the mask R-CNN preening behavior detector could be a useful tool to automatically identify preening behaviors of individual hens in group settings.