臺灣與亞洲養豬產業為重要之動物性蛋白來源，然近年面臨人力短缺與高齡化等問題，使得分娩監測高度仰賴人工長時間觀察之作業模式逐漸難以維持。分娩過程中若無法即時察覺關鍵事件，可能導致難產或仔豬壓迫等風險，進而影響生產效率與經濟效益。因此，本研究針對上述問題，致力於建立一套可於實際畜舍環境中長時間運作之自動化分娩監測方法，以提升監測效率並降低相關風險。
本研究場域為實際分娩舍，母豬於接近分娩期間移入分娩床接受監測，固定式攝影機則由分娩床上方或斜上方持續觀測分娩床與產道區域。基於此固定視角之監視影像，本研究將分娩過程中之「露頭事件」界定為逐影格二元分類問題，判斷產道區域是否出現新生仔豬從母豬產道露頭情形。在此設定下，本文採用輕量化卷積神經網路（Convolutional Neural Network, CNN）作為基礎架構，透過整合不同層級之影像特徵與強化關鍵區域資訊，使模型能在複雜背景與細微線索下穩定辨識露頭事件。該方法在有限運算資源條件下，兼顧辨識效能與推論效率，使其適用於畜舍邊緣設備之長時間即時監測需求。
實驗結果顯示，所提出方法在分娩影像辨識任務中具有良好表現，其 Accuracy 與 F1-score 分別達 69.98% 與 72.60%，並在維持約 1.32 ms 推論延遲與低模型大小之條件下，提升整體辨識穩定性。結果顯示，本研究方法能在實務場域限制下，有效輔助分娩監測，降低人力負擔並提升即時反應能力。

Timely recognition of sow farrowing events is important in precision livestock farming, but practical monitoring systems often rely on fixed cameras and limited computational resources. This study formulates the visible emergence of a newborn piglet from the birth canal as frame-level binary classification between Exposed and Not Exposed. To address small target regions, occlusion, and illumination variation in surveillance images, we propose MSFU-Net, a lightweight SqueezeNet-based model with multi-scale feature fusion, bidirectional query scoring, gated residual fusion, and guided pooling. On 48,000 labeled images from 8 sows in a real farrowing house, MSFU-Net achieved 69.98% Accuracy, 72.60% F1-score, and 0.7565 AUC under leave-one-group-out cross-validation, with 1.32 ms latency and a 9.42 MB model size. These results show that MSFU-Net balances recognition performance and computational efficiency for fixed-view sow farrowing monitoring.

Bochkovskiy, A., Wang, C.-Y., and Liao, H.-Y. M. (2020). Yolov4: Optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934.

Box, G. E. P., Jenkins, G. M., Reinsel, G. C., and Ljung, G. M. (2015). Time Series Analysis: Forecasting and Control. John Wiley & Sons, 5th edition.

Chen, B.-L., Cheng, T.-H., Huang, Y.-C., Hsieh, Y.-L., Hsu, H.-C., Lu, C.-Y., Huang, M.-H., Nien, S.-Y., and Kuo, Y.-F. (2023a). Developing an automatic warning system for anomalous chicken dispersion and movement using deep learning and machine learning. Poultry Science, 102(12):103040.

Chen, J., Zhou, J., Liu, L., Shu, C., Shen, M., and Yao, W. (2023b). Sow farrowing early warning and supervision for embedded board implementations. Sensors, 23(2):727.

Cubuk, E. D., Zoph, B., Mane, D., Vasudevan, V., and Le, Q. V. (2019). Autoaugment: Learning augmentation policies from data. In CVPR.

Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L. (2009). Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. IEEE.

Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020). An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929.

Ge, Z., Liu, S., Wang, F., Li, Z., and Sun, J. (2021). Yolox: Exceeding yolo series in 2021. arXiv preprint arXiv:2107.08430.

Ghosh, A., Chakraborty, D., and Law, A. (2018). Artificial intelligence in internet of things. CAAI Transactions on Intelligence Technology, 3(4):208–218.

González-Baldizón, Y., Pérez-Patricio, M., Camas-Anzueto, J. L., Rodríguez-Elías, O. M., Escobar-Gómez, E. N., Vazquez-Delgado, H. D., Guzman-Rabasa, J. A., and Fragoso-Mandujano, J. A. (2022). Lamb behaviors analysis using a predictive cnn model and a single camera. Applied Sciences, 12(9):4712.

Goodfellow, I., Bengio, Y., and Courville, A. (2016). Deep Learning. MIT Press.

He, K., Zhang, X., Ren, S., and Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778.

Hochreiter, S. and Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8):1735–1780.

Hu, J., Shen, L., and Sun, G. (2018). Squeeze-and-excitation networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 7132–7141.

Huang, J., Rathod, V., Sun, C., Zhu, M., Korattikara, A., Fathi, A., Fischer, I., Wojna, Z., Song, Y., Guadarrama, S., and Murphy, K. (2017). Speed/accuracy trade-offs for modern convolutional object detectors. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

Iandola, F. N., Han, S., Moskewicz, M. W., Ashraf, K., Dally, W. J., and Keutzer, K. (2016). Squeezenet: Alexnet-level accuracy with 50x fewer parameters and <0.5mb model size. arXiv preprint arXiv:1602.07360.

Jocher, G. (2020). YOLOv5 by Ultralytics.

Krizhevsky, A., Sutskever, I., and Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. In Advances in Neural Information Processing Systems (NeurIPS), volume 25, pages 1097–1105. Curran Associates, Inc.

Küster, S., Nolte, P., Meckbach, C., Stock, B., and Traulsen, I. (2021). Automatic behavior and posture detection of sows in loose farrowing pens based on 2d-video images. Frontiers in Animal Science, 2:758165.

Le Dividich, J. and Herpin, P. (1994). Effects of climatic conditions on the performance, metabolism and health status of weaned piglets: a review. Livestock Production Science, 38(2):79–90.

Lipper, L., Thornton, P., Campbell, B. M., Baedeker, T., Braimoh, A., Bwalya, M., Caron, P., Cattaneo, A., Garrity, D., Henry, K., McCarthy, N., Zilberman, D., et al. (2014). Climate-smart agriculture for food security. Nature Climate Change, 4(12):1068–1072.

Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.-Y., and Berg, A. C. (2016). SSD: Single shot multibox detector. In ECCV.

Loshchilov, I. and Hutter, F. (2017). Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101.

McDonagh, J., Tzimiropoulos, G., Slinger, K. R., Huggett, Z. J., Down, P. M., and Bell, M. J. (2021). Detecting dairy cow behavior using vision technology. Agriculture, 11(7):675.

Mluba, H. S., Atif, O., Lee, J., Park, D., and Chung, Y. (2024). Pattern mining-based pig behavior analysis for health and welfare monitoring. Sensors, 24(7):2185.

NVIDIA Corporation (2021). TensorRT: High Performance Deep Learning Inference Optimizer and Runtime. https://developer.nvidia.com/tensorrt.

Oh, S.-H. and Whitley, N. C. (2011). Pork production in china, japan and south korea. Asian-Australasian Journal of Animal Sciences, 24(11):1629–1636.

Oktay, O., Schlemper, J., Le Folgoc, L., Lee, M., Heinrich, M., Misawa, K., Mori, K., McDonagh, S., Hammerla, N. Y., Kainz, B., Glocker, B., and Rueckert, D. (2018). Attention u-net: Learning where to look for the pancreas. arXiv preprint arXiv:1804.03999.

Redmon, J. and Farhadi, A. (2018). Yolov3: An incremental improvement.

Ronneberger, O., Fischer, P., and Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, volume 9351 of Lecture Notes in Computer Science, pages 234–241, Cham. Springer.

Schuster, M. and Paliwal, K. K. (1997). Bidirectional recurrent neural networks. IEEE Transactions on Signal Processing, 45(11):2673–2681.

Song, S., Kang, T., Lim, K., Kim, K., and Yi, H. (2024). Sow posture analysis and estrus prediction using closed-circuit television cameras. IEEE Access, 12:17460–17466.

Tan, M., Chen, B., Pang, R., Vasudevan, V., Sandler, M., Howard, A., and Le, Q. V. (2019). Mnasnet: Platform-aware neural architecture search for mobile. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 2820–2828.

Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., and Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 30.

Wang, C.-Y., Bochkovskiy, A., and Liao, H.-Y. M. (2023). Yolov7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 7464–7475.

Wang, X., Girshick, R., Gupta, A., and He, K. (2018). Non-local neural networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 7794–7803.

Wojke, N., Bewley, A., and Paulus, D. (2017). Simple online and realtime tracking with a deep association metric. In 2017 IEEE International Conference on Image Processing (ICIP), pages 3645–3649. IEEE.

Woo, S., Park, J., Lee, J., and Kweon, I. S. (2018). Cbam: Convolutional block attention module. In Proceedings of the European Conference on Computer Vision (ECCV), pages 3–19.

Zambon, I., Cecchini, M., Egidi, G., Saporito, M. G., and Colantoni, A. (2019). Revolution 4.0: Industry vs. agriculture in a future development for smes. Processes, 7(1):36.

Zhang, H., Goodfellow, I., Metaxas, D., and Odena, A. (2019). Self-attention generative adversarial networks. In International conference on machine learning, pages 7354–7363. PMLR.

Zhang, Y., Sun, P., Jiang, Y., Yu, T.-Y., Weng, C.-Y., Yuan, L., and Luo, P. (2022). Bytetrack: Multi-object tracking by associating every detection box. In European Conference on Computer Vision (ECCV), pages 1–21. Springer.

陳晉茹 (2024). 使用多尺度注意力網路進行鋼鐵表面瑕疵辨識. 碩士論文, 國立成功大學. 臺灣博碩士論文知識加值系統.