Sheep detection and segmentation will play a crucial role in promoting the implementation of precision livestock farming in the future. In sheep farms, the characteristics of sheep that have the tendency to congregate and irregular contours cause difficulties for computer vision tasks, such as individual identification, behavior recognition, and weight estimation of sheep. Sheep instance segmentation is one of the methods that can mitigate the difficulties associated with locating and extracting different individuals from the same category. To improve the accuracy of extracting individual sheep locations and contours in the case of multiple sheep overlap, this paper proposed two-stage sheep instance segmentation SheepInst based on the Mask R-CNN framework, more specifically, RefineMask. Firstly, an improved backbone network ConvNeXt-E was proposed to extract sheep features. Secondly, we improved the structure of the two-stage object detector Dynamic R-CNN to precisely locate highly overlapping sheep. Finally, we enhanced the segmentation network of RefineMask by adding spatial attention modules to accurately segment irregular contours of sheep. SheepInst achieves 89.1%, 91.3%, and 79.5% in box AP, mask AP, and boundary AP metric on the test set, respectively. The extensive experiments show that SheepInst is more suitable for sheep instance segmentation and has excellent performance.

The behavior of pigs is intricately tied to their health status, highlighting the critical importance of accurately recognizing pig behavior, particularly abnormal behavior, for effective health monitoring and management. This study addresses the challenge of accommodating frequent non-rigid deformations in pig behavior using deformable convolutional networks (DCN) to extract more comprehensive features by incorporating offsets during training. To overcome the inherent limitations of traditional DCN offset weight calculations, the study introduces the multi-path coordinate attention (MPCA) mechanism to enhance the optimization of the DCN offset weight calculation within the designed DCN-MPCA module, further integrated into the cross-scale cross-feature (C2f) module of the backbone network. This optimized C2f-DM module significantly enhances feature extraction capabilities. Additionally, a gather-and-distribute (GD) mechanism is employed in the neck to improve non-adjacent layer feature fusion in the YOLOv8 network. Consequently, the novel DM-GD-YOLO model proposed in this study is evaluated on a self-built dataset comprising 11,999 images obtained from an online monitoring platform focusing on pigs aged between 70 and 150 days. The results show that DM-GD-YOLO can simultaneously recognize four common behaviors and three abnormal behaviors, achieving a precision of 88.2%, recall of 92.2%, and mean average precision (mAP) of 95.3% with 6.0MB Parameters and 10.0G FLOPs. Overall, the model outperforms popular models such as Faster R-CNN, EfficientDet, YOLOv7, and YOLOv8 in monitoring pens with about 30 pigs, providing technical support for the intelligent management and welfare-focused breeding of pigs while advancing the transformation and modernization of the pig industry.

Livestock behavior serves as a crucial indicator of physiological health. Leveraging deep learning techniques to automatically recognize dairy goat behaviors, particularly abnormal ones, enables early detection of potential health and environmental issues. To address the challenges of recognizing small-target behaviors in complex environments, a multi-scale and lightweight behavior recognition model for dairy goats called GSCW-YOLO was proposed. The model integrates Gaussian Context Transformation (GCT) and the Content-Aware Reassembly of Features (CARAFE) upsampling operator, enhancing the YOLOv8n framework’s attention to behavioral features, reducing interferences from complex backgrounds, and improving the ability to distinguish subtle behavior differences. Additionally, GSCW-YOLO incorporates a small-target detection layer and optimizes the Wise-IoU loss function, increasing its effectiveness in detecting distant small-target behaviors and transient abnormal behaviors in surveillance videos. Data for this study were collected via video surveillance under varying lighting conditions and evaluated on a self-constructed dataset comprising 9213 images. Experimental results demonstrated that the GSCW-YOLO model achieved a precision of 93.5%, a recall of 94.1%, and a mean Average Precision (mAP) of 97.5%, representing improvements of 3, 3.1, and 2 percentage points, respectively, compared to the YOLOv8n model. Furthermore, GSCW-YOLO is highly efficient, with a model size of just 5.9 MB and a frame per second (FPS) of 175. It outperforms popular models such as CenterNet, EfficientDet, and other YOLO-series networks, providing significant technical support for the intelligent management and welfare-focused breeding of dairy goats, thus advancing the modernization of the dairy goat industry.

Mining the correlation has attracted widespread attention in the research community because of its advantages in understanding the dependencies between objects. In this paper, a correlated graph pattern searching scheme has been proposed, that is, provided with a query g as a structured pattern (i.e., a graph), our algorithm is capable of retrieving the top-k graphs that most likely correlated with g. Traditional methods treat graph streams as static records, which is computational infeasible or ineffective because of the complexity of searching correlated patterns in a dynamic graph stream. In this paper, by relying on sliding windows to separate graph streams in chunks, we propose a Hoe-PGPL algorithm to handle the top-k correlated patterns searching from a dynamic perspective. Our algorithm applies Hoeffding bound and two-level (Sliding window level and local batch level) candidate inspection to discover potential graph candidates and determine the similarity of these candidates without double-checking the previous stream. Theoretical analysis shows that our method can guarantee the quality of the returned answers, and our experiments also present that Hoe-PGPL has an excellent performance with aspects of precision, recall, runtime, and resource consumption.