Abstract
The individual identification of cattle is crucial for the efficient management of large farms, and advanced identification technologies that can monitor cattle behavior in real time are essential for increasing agricultural efficiency, promoting the digital transformation of animal husbandry, and improving animal welfare. This paper introduces a novel network called OP Mask R-CNN for individual cattle identification, which combines Open Pose with the Mask R-CNN network. We present three key strategies to improve the identification of individual cattle. First, we optimize the number of convolutional layers in the Mask R-CNN backbone network, i.e., ResNet101. Second, we introduce an Open Pose-based bovine skeleton feature extraction method. Finally, we construct a fusion mechanism that combines the attention module, the convolutional block attention module (CBAM), the open pose module, and the ResNet101. This work strikes a balance between accuracy and complexity that supports the development of a lightweight bovine individual recognition technique.