中藥材的使用是中醫當中很重要的療法之一。現今相關的中藥材研究都是需要一些儀器去分析，加上並沒有很深入探討易混淆中藥的辨識。一般使用者還是只能透過口嘗、鼻聞、外觀來辨別中藥材。近年來，深度學習被廣泛應用於圖像識別，然而這項技術並沒有在中藥辨識有很完整的發展。因此，在這篇論文中我們提出了一個使用階層式分群之卷積神經網絡在易混淆中藥影像的辨識系統，透過資料分群讓卷積神經網絡能對易混淆中藥萃取出更具代表性的特徵。在實驗部分，我們透過自己拍攝的24種中藥(其中包含10種易混淆組別)進行評估，實驗證實卷積神經網絡的方法比其他傳統方法的辨識能力更好，有用階層式的分群能達到更高的辨識度。我們還探討了4種不同手智慧型手機對辨識系統的影響，仍然有不錯的結果，並加入不同手機的照片以及資料擴增的方法來增加訓練資料去提升辨識度。最後，我們將此系統建立在伺服器上，使用者只要透過手機拍攝並且上傳到我們的系統進行辨識，就可以收到此中藥的相關資訊。

Using of Chinese Herbal Medicines (CHMs) plays an important role of treatment in Traditional Chinese medicine (TCM). Many CHMs researches require some instrument to analyze. In addition, they are not in-depth discussion the recognition of easy-confused herbs. Recently, deep learning method has widely used in image recognition. However, this technology is not complete development in the CHMs recognition yet.
In this paper, we propose a vision-based CHMs recognition system using hierarchical clustering convolutional neural networks(CNN). CNN can be used to extract more representative features of easily-confused herbs by data clustering.
In experimental part, our recognition method is evaluated by 24 kinds of herbs which collected by ourselves (Containing 10 easy-confused herbs pairs). Experiments show that CNN method is better than other hand-crafted methods, and it can achieve higher accuracy rates using hierarchical clustering. We also explore the impact of four different smartphones on the recognition system, and still get acceptable results. We add other smartphones images and data augmentation method to increase training data to improve this problem.
Finally, we build a TCM herbs recognition system on the server, and the users can receive information about the herbs which they have them as long as they capture images by smartphones and upload to our system.

[1] Sebastian Bach, Alexander Binder, Grégoire Montavon, Frederick Klauschen, KlausRobert Müller, and Wojciech Samek. On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance propagation. PloS one, 10(7):e0130140, 2015.
[2] Jake Bouvrie. Notes on convolutional neural networks. 2006.
[3] Chih-Chung Chang and Chih-Jen Lin. Libsvm: a library for support vector machines. ACM transactions on intelligent systems and technology (TIST), 2(3):27, 2011.
[4] Hao Chen, Josiah Poon, Simon Poon, Kelvin Chan, Ka Ho Wong, and Daniel Sze. Evaluation of herbal chromatographie fingerprint identification systems. In Bioinformatics and Biomedicine Workshops (BIBMW), 2012 IEEE International Conference on, pages 418–423. IEEE, 2012.
[5] Navneet Dalal and Bill Triggs. Histograms of oriented gradients for human detection. In Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on, volume 1, pages 886–893. IEEE, 2005.
[6] Luo Dehan, Wang Jia, Chen Yimin, and GholamHosseini Hamid. Classification of chinese herbal medicines based on svm. In Information Science, Electronics and Electrical Engineering (ISEEE), 2014 International Conference on, volume 1, pages 453–456. IEEE, 2014.
[7] Brendan J Frey and Delbert Dueck. Clustering by passing messages between data points. science, 315(5814):972–976, 2007.
[8] YL Ho, YS Chang, and IH Lin. Illustration of commonly misused chinese crude drug species in taiwan. 2006.
[9] Yangqing Jia, Evan Shelhamer, Jeff Donahue, Sergey Karayev, Jonathan Long, Ross Girshick, Sergio Guadarrama, and Trevor Darrell. Caffe: Convolutional architecture for fast feature embedding. In Proceedings of the 22nd ACM international conference on Multimedia, pages 675–678. ACM, 2014.
[10] Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems, pages 1097–1105, 2012.
[11] Yann LeCun, Léon Bottou, Yoshua Bengio, and Patrick Haffner. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11):2278–2324, 1998.
[12] Z Li. Chinese herbal medicine feature extraction and identification system. 2013.
[13] Changjiang Liu, Xuling Wu, and Wei Xiong. Chinese herbal medicine classification based on bp neural network. Journal of Software, 9(4):938–944, 2014.
[14] David G Lowe. Distinctive image features from scale-invariant keypoints. International journal of computer vision, 60(2):91–110, 2004.
[15] Dehan Luo, Danjun Fan, Hao Yu, and Zhimin Li. A new processing technique for the identification of chinese herbal medicine. In Computational and Information Sciences (ICCIS), 2013 Fifth International Conference on, pages 474–477. IEEE, 2013.
[16] James MacQueen et al. Some methods for classification and analysis of multivariate observations. In Proceedings of the fifth Berkeley symposium on mathematical statistics and probability, volume 1, pages 281–297. Oakland, CA, USA., 1967.
[17] Xuehong Mao, Samer Hijazi, Raúl Casas, Piyush Kaul, Rishi Kumar, and Chris Rowen. Hierarchical cnn for traffic sign recognition. In Intelligent Vehicles Symposium (IV), 2016 IEEE, pages 130–135. IEEE, 2016.
[18] Andrew Y Ng, Michael I Jordan, and Yair Weiss. On spectral clustering: Analysis and an algorithm. In Advances in neural information processing systems, pages 849–856, 2002.
[19] Maria-Elena Nilsback and Andrew Zisserman. Automated flower classification over a large number of classes. In Computer Vision, Graphics & Image Processing, 2008. ICVGIP’08. Sixth Indian Conference on, pages 722–729. IEEE, 2008.
[20] Timo Ojala, Matti Pietikainen, and Topi Maenpaa. Multiresolution gray-scale and rotation invariant texture classification with local binary patterns. IEEE Transactions on pattern analysis and machine intelligence, 24(7):971–987, 2002.
[21] Xiao-Ming Ren, Xiao-Feng Wang, and Yang Zhao. An efficient multi-scale overlapped block lbp approach for leaf image recognition. In International Conference on Intelligent Computing, pages 237–243. Springer, 2012.
[22] David E Rumelhart, Geoffrey E Hinton, and Ronald J Williams. Learning representations by back-propagating errors. Cognitive modeling, 5(3):1, 1988.
[23] Karen Simonyan and Andrew Zisserman. Very deep convolutional networks for largescale image recognition. arXiv preprint arXiv:1409.1556, 2014.
[24] Xin Sun and Huinan Qian. Chinese herbal medicine image recognition and retrieval by convolutional neural network. PloS one, 11(6):e0156327, 2016.
[25] Ou Tao, Zhaozhou Lin, XianBao Zhang, Yun Wang, and Yanjiang Qiao. Research on identification model of chinese herbal medicine by texture feature parameter of transverse section image. World Science and Technology-Modernization of Traditional Chinese Medicine and Materia Medica, pages 2558–2562, 2014.
[26] Ou Tao, Yanling Zhang, Qian Chen, Yun Wang, and Yanjiang Qiao. Extraction of texture feature parameter of transverse section in chinese herbal medicine by gray-level co-occurrence matrix. World Science and Technology-Modernization of Traditional Chinese Medicine and Materia Medica, pages 2531–2537, 2014.
[27] Jing Wang, Hongwei Kong, Zimin Yuan, Peng Gao, Weidong Dai, Chunxiu Hu, Xin Lu, and Guowang Xu. A novel strategy to evaluate the quality of traditional Chinese medicine based on the correlation analysis of chemical fingerprint and biological effect. Journal of pharmaceutical and biomedical analysis, 83:57–64, 2013.
[28] Yiwen Wu. Global smartphone user penetration forecast by 88 countries : 2007-2022. 2016.
[29] Xue-Yang Xiao, Rongxiang Hu, Shan-Wen Zhang, and Xiao-Feng Wang. Hog-based approach for leaf classification. In Advanced Intelligent Computing Theories and Applications. With Aspects of Artificial Intelligence, pages 149–155. Springer, 2010.
[30] Zhicheng Yan, Hao Zhang, Robinson Piramuthu, Vignesh Jagadeesh, Dennis DeCoste, Wei Di, and Yizhou Yu. Hd-cnn: Hierarchical deep convolutional neural network for large scale visual recognition. arXiv preprint arXiv:1410.0736, 2014.
[31] Hossam M Zawbaa, Mona Abbass, Sameh H Basha, Maryam Hazman, and Abul Ella Hassenian. An automatic flower classification approach using machine learning algorithms. In Advances in Computing, Communications and Informatics (ICACCI, 2014 International Conference on, pages 895–901. IEEE, 2014.