Individual identification of dairy cows based on convolutional neural networks

Individual identification of each cow is significant for precision livestock farming. In this paper, we propose a novel contactless cow identification method based on convolutional neural networks. We first collected a set of side-view images of dairy cows, then employed the YOLO model to detect the cow object in the side-view image, and finally fine-tuned a convolutional neural network model to classify each individual cow. In our experiments, a total of 105 side-view images of cows were collected, and the proposed method achieved an accuracy of 96.65% in cow identification, which outperformed existing experiments. Experimental results demonstrate the effectiveness of the proposed method for cow identification and the potential for our method to be applied to other livestock.