Deep learning algorithms to identify individual finishing pigs using 3D data

Authors: PAUDEL, SHIVA - University Of Nebraska; BROWN-BRANDL, TAMI - University Of Nebraska; Rohrer, Gary; SHARMA, SUDHENDU RAJ - University Of Nebraska

Submitted to: Biosystems Engineering
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/3/2025
Publication Date: 7/1/2025
Citation: Paudel, S., Brown-Brandl, T.M., Rohrer, G.A., Sharma, S.R. 2025. Deep learning algorithms to identify individual finishing pigs using 3D data. Biosystems Engineering. 255. Article 104143. https://doi.org/10.1016/j.biosystemseng.2025.104143.
DOI: https://doi.org/10.1016/j.biosystemseng.2025.104143

Interpretive Summary: In order to truly create precision livestock farming devices, individual animals need to be identified. While ear tags are a simple solution, they add costs in grow-finish pigs and require labor for application. Therefore, this study aims to investigate pig identification based on machine-learning from a 3D depth camera mounted over the waterer. Three different stages of grow-finish pigs were tested; early (first 15 days), mid-finishing (days 45-60) and late finishing (days 78-91). The procedure was tested to see how well this method would perform for short-term (less than 15 days) and long-term identification (30-90 days) between different stages of growth. For this study, 26 animals were continuously recorded. The first 4 days of images from each stage of growth were used to develop a computer model and the model was then used on images from subsequent days. The model developed from the early growth stage could not achieve an acceptable accuracy of identification for any phase of the study. However, models developed from the mid- and late-finishing stages had accuracies of at least 0.68. The model developed from mid-finishing animals was able to identify pigs through the final stages of finishing with an accuracy of 0.47 and could achieve an accuracy of 0.89 if the model was retrained halfway through the final period. This study determined that computer vision using machine-learning on 3D depth camera data can accurately identify grow-finish pigs after the mid-growth stage (approximately 100 days of age), but the model requires retraining with a small volume of data between the mid and late finishing stage to achieve the highest accuracy. Utilizing a computer image system to individually identify pigs in a commercial grow-finish barn would reduce cost of production and enable new methods to monitor animal well-being more thoroughly throughout this phase of production.

Technical Abstract: The application of precision livestock farming technology is heavily reliant on the identification of individuals. However, due to the cost and time constraints, finishing pigs are rarely tagged or otherwise identified. Therefore, the objectives of this study were to determine the feasibility of using deep learning on 3D spatial data to identify individual finishing pigs and to quantify the amount of data required, image resolution needed, and frequency of retraining for continuous identification using two different architectures: PointNet (which utilises point clouds directly) and 3D convolution neural network (3D CNN). Digital/depth images were collected using ToF (Time of Flight) camera positioned over RFID (Radio Frequency Identification) instrumented drinkers. A subset of this data were used for this initial validation study, which included 31976 images from eight pigs over 14 days. The data were then processed to create different sets of training and testing data with varying point sets (1500, 3000, 6000, 12000, 24000, and 48000) for point clouds and voxel sizes (50, 35, 25, and 15 mm) for 3D CNN. The findings revealed that the 3D CNN model achieved the highest F1 score of 0.91 after the sixth training session with a point voxel size of 15 mm. PointNet achieved its highest F1 score of 0.90 after five training sessions with a point set size of 1500 points. This study underscores the potential of utilising deep learning techniques for the purpose of individual pig identification within actual barn environments, including those with natural lighting conditions.