本研究聚焦於濃縮中藥製劑（Concentrated Chinese Medicines, CCMs）的組成分析。此類製劑在臨床與市售應用中日益普及，引發了對其成分組成與品質一致性的關注。由於製程中常混合多種製成方法的藥材粉末與賦形劑（如澱粉），使得成分鑑別與來源溯源更加困難。高效液相層析（High-Performance Liquid Chromatography, HPLC）是監測與鑑定中藥成分的重要技術，然而其高昂的分析成本與冗長的流程限制了人工智慧模型大量訓練資料的取得。為解決此問題，本研究提出一套基於人工智慧與點雲（Point Clouds）的層次式分類架構，用以分析小規模 HPLC 數據。透過將層析訊號轉換為三維點雲形式，進行模型訓練，使模型能進行多層級分類，以鑑別未知 CCM 樣本中的藥材種類並推估其粉末與浸膏比例。實驗結果顯示，該模型即使在樣本有限的條件下仍展現穩健的分類表現，並經由資料擴增與消融實驗驗證其穩定性。研究結果證實，在少量 HPLC 數據條件下進行中藥成分結構推論是可行的，並為未來智慧製藥與品質管控提供技術基礎。

This study focuses on the compositional analysis of Concentrated Chinese Medicines (CCMs), which have become increasingly prevalent in both clinical and commercial applications, raising concerns about their compositional consistency and quality stability. During manufacturing, CCMs are often produced by blending herbal powders prepared through different extraction methods together with excipients such as starch, which complicates chemical identification and source traceability. High-performance liquid chromatography (HPLC) serves as a key analytical technique for monitoring and characterizing herbal components; however, its high analytical cost and lengthy process restrict the acquisition of large-scale training data for artificial intelligence (AI) models. To address this issue, this study proposes a hierarchical classification framework based on AI and point cloud representation to analyze small-scale HPLC data. By transforming chromatographic signals into three-dimensional point clouds, the model is trained to perform multi-level classification for identifying herbal species and estimating the powder-to-extract ratios in unknown CCM samples. Experimental results demonstrate that the proposed model exhibits robust classification performance even under limited data conditions and maintains stability through data augmentation and ablation studies. The findings confirm the feasibility of structural inference of CCM compositions from small HPLC datasets and provide a technical foundation for future intelligent pharmaceutical manufacturing and quality control applications.

Cheng, H., Zhang, Y., Wang, X., Li, W., & Sun, J. (2022). Establishment of anthocyanin fingerprint in black wolfberry fruit for quality and geographical origin identification. LWT – Food Science and Technology, 157, 113080.
Grilli, E., Menna, F., & Remondino, F. (2017). A review of point clouds segmentation and classification algorithms. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 42, 339–344.
Guo, M.-H., Cai, J.-X., Liu, Z.-N., Mu, T.-J., Martin, R. R., & Hu, S.-M. (2021). PCT: Point cloud transformer. Computational Visual Media, 7(2), 187–199.
Han, J., Li, Q., Zhang, S., Chen, W., Zhou, X., & Zhao, Y. (2022). Qualitative and quantitative evaluation of Flos Puerariae by using chemical fingerprint in combination with chemometrics method. Journal of Pharmaceutical Analysis, 12(3), 489–499.
Jain, D. S., Shrivastava, K. S., Agrawal, A., & Gupta, K. A. (2024). WHO guidelines for quality control of herbal medicines: From cultivation to consumption. International Journal of Pharmaceutical Chemistry and Analysis, 212–225.
Li, Y., Bu, R., Sun, M., Wu, W., Di, X., & Chen, B. (2018). PointCNN: Convolution on X-transformed points. Advances in Neural Information Processing Systems, 31, 828–838.
Luo, H., Li, Q., Flower, A., Lewith, G., & Liu, J. (2012). Comparison of effectiveness and safety between granules and decoction of Chinese herbal medicine: A systematic review of randomized clinical trials. Journal of Ethnopharmacology, 140(3), 555–567.
Maturana, D., & Scherer, S. (2015). VoxNet: A 3D convolutional neural network for real-time object recognition. In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (pp. 922–928). IEEE.
Qi, C. R., Su, H., Mo, K., & Guibas, L. J. (2017). PointNet: Deep learning on point sets for 3D classification and segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 77–85). IEEE Computer Society.
Qi, C. R., Yi, L., Su, H., & Guibas, L. J. (2017). PointNet++: Deep hierarchical feature learning on point sets in a metric space. Advances in Neural Information Processing Systems, 5099–5108.
Shorten, C., & Khoshgoftaar, T. M. (2019). A survey on image data augmentation for deep learning. Journal of Big Data, 6(1), 60.
Si, H., & Wei, X. (2024). Feature extraction and representation learning of 3D point cloud data. Image and Vision Computing, 142, 104890.
Wang, Y., Sun, Y., Liu, Z., Sarma, S. E., Bronstein, M. M., & Solomon, J. M. (2019). Dynamic graph CNN for learning on point clouds. ACM Transactions on Graphics, 38(5), 1–12.
Welfare of Health and Ministry. (2022). Taiwan herbal pharmacopeia (4th ed.). Taipei: Ministry of Health and Welfare.
Welfare of Health and Ministry. (2023). Guidelines for drug registration and review. Taipei: Ministry of Health and Welfare.
Wu, W., Qi, Z., & Li, F. (2019). PointConv: Deep convolutional networks on 3D point clouds. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 9621–9630). IEEE.
Xie, P., Chen, S., Liang, Y. Z., Wang, X., Tian, R., & Upton, R. (2006). Chromatographic fingerprint analysis: A rational approach for quality assessment of traditional Chinese herbal medicine. Journal of Chromatography A, 1112(1–2), 171–180.
Yin, T., Zhou, X., & Krähenbühl, P. (2021). Multimodal virtual point 3D detection. Advances in Neural Information Processing Systems, 34, 16494–16507.
Zhang, H., Liu, Y., Chen, J., Wu, T., & Li, X. (2023). Deep learning-based 3D point cloud classification: A systematic survey and outlook. Displays, 79, 102456.
Zhang, X., Wang, Q., Zhang, J., & Zhong, Z. (2019). Adversarial AutoAugment. arXiv preprint arXiv:1912.11188.
Zhu, Q., Fan, L., & Weng, N. (2024). Advancements in point cloud data augmentation for deep learning: A survey. Pattern Recognition, 153, 110532.
Cai, Q., Pan, Y., Yao, T., Yan, C., & Mei, T. (2018). Memory matching networks for one-shot image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 4080–4088).
Finn, C., Abbeel, P., & Levine, S. (2017). Model-agnostic meta-learning for fast adaptation of deep networks. In Proceedings of the International Conference on Machine Learning (pp. 1126–1135). PMLR.
Koch, G., Zemel, R., & Salakhutdinov, R. (2015). Siamese neural networks for one-shot image recognition. In ICML Deep Learning Workshop (Vol. 2, No. 1, pp. 1–30). Lille, France.
Li, H., Eigen, D., Dodge, S., Zeiler, M., & Wang, X. (2019). Finding task-relevant features for few-shot learning by category traversal. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 1–10).
Mishra, N., Rohaninejad, M., Chen, X., & Abbeel, P. (2017). A simple neural attentive meta-learner. arXiv preprint arXiv:1707.03141.
Munkhdalai, T., & Yu, H. (2017). Meta networks. In Proceedings of the International Conference on Machine Learning (pp. 2554–2563). PMLR.
Munkhdalai, T., Yuan, X., Mehri, S., & Trischler, A. (2018). Rapid adaptation with conditionally shifted neurons. In Proceedings of the International Conference on Machine Learning (pp. 3664–3673). PMLR.
Oreshkin, B., Rodríguez López, P., & Lacoste, A. (2018). TADAM: Task dependent adaptive metric for improved few-shot learning. Advances in Neural Information Processing Systems, 31.
Rusu, A. A., et al. (2018). Meta-learning with latent embedding optimization. arXiv preprint arXiv:1807.05960.
Santoro, A., Bartunov, S., Botvinick, M., Wierstra, D., & Lillicrap, T. (2016). Meta-learning with memory-augmented neural networks. In Proceedings of the International Conference on Machine Learning (pp. 1842–1850). PMLR.
Snell, J., Swersky, K., & Zemel, R. (2017). Prototypical networks for few-shot learning. Advances in Neural Information Processing Systems, 30.
Sun, Q., Liu, Y., Chua, T.-S., & Schiele, B. (2019). Meta-transfer learning for few-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 403–412).
Sung, F., Yang, Y., Zhang, L., Xiang, T., Torr, P. H., & Hospedales, T. M. (2018). Learning to compare: Relation network for few-shot learning. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 1199–1208).
Vinyals, O., Blundell, C., Lillicrap, T., & Wierstra, D. (2016). Matching networks for one shot learning. Advances in Neural Information Processing Systems, 29.
Wang, Z., et al. (2024). A comprehensive survey on data augmentation. arXiv preprint arXiv:2405.09591.
Xu, M., Yoon, S., Fuentes, A., & Park, D. S. (2023). A comprehensive survey of image augmentation techniques for deep learning. Pattern Recognition, 137, 109347.
Yoon, S. W., Seo, J., & Moon, J. (2019). TapNet: Neural network augmented with task-adaptive projection for few-shot learning. In Proceedings of the International Conference on Machine Learning (pp. 7115–7123). PMLR.
Zhang, Y., Gong, M., Li, J., Feng, K., & Zhang, M. (2024). Few-shot learning with enhancements to data augmentation and feature extraction. IEEE Transactions on Neural Networks and Learning Systems.