2.2.1. Dataset Formatting

Different sow objects existed in an image; thus, the classification results of the target sows would be considerably perturbed. The original image was clipped to ensure the accuracy of the classification results. The image was trimmed as a rectangular box with coordinates of (500, 0), (1250, 0), (0, 1080), and (500, 1080). Subsequently, the images were divided into training, validation, and test sets in a ratio of 8:1:1. The training set was used to train the BCS model; the validation set was used to adjust the parameters and select the optimal model; and the test set was used to assess the performance of the sow BCS model [6].