Terahertz Spectroscopic Identification with Deep Belief Network
MA Shuai1, SHEN Tao1, 2*, WANG Rui-qi1, LAI Hua1, YU Zheng-tao1
1. School of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650500, China 2. School of Materials Science and Engineering, Kunming University of Science and Technology, Kunming 650500, China
Abstract：Feature extraction and classification are the key issues of terahertz spectroscopy identification. Because many materials have no apparent absorption peaks in the terahertz band, it is difficult to extract theirs terahertz spectroscopy feature and identify. To this end, a novel of identify terahertz spectroscopy approach with Deep Belief Network (DBN) was studied in this paper, which combines the advantages of DBN and K-Nearest Neighbors (KNN) classifier. Firstly, cubic spline interpolation and S-G filter were used to normalize the eight kinds of substances (ATP, Acetylcholine_Bromide, Bifenthrin, Buprofezin, Carbazole, Bleomycin, Buckminster and Cylotriphosphazene) terahertz transmission spectra in the range of 0.9～6 THz. Secondly, the DBN model was built by two restricted Boltzmann machine (RBM) and then trained layer by layer using unsupervised approach. Instead of using handmade features, the DBN was employed to learn suitable features automatically with raw input data. Finally, a KNN classifier was applied to identify the terahertz spectrum. Experimental results show that using the feature learned by DBN can identify the terahertz spectrum of different substances with the recognition rate of over 90%, which demonstrates that the proposed method can automatically extract the effective features of terahertz spectrum. Furthermore, this KNN classifier was compared with others (BP neural network, SOM neural network and RBF neural network). Comparisons showed that the recognition rate of KNN classifier is better than the other three classifiers. Using the approach that automatic extract terahertz spectrum features by DBN can greatly reduce the workload of feature extraction. This proposed method shows a promising future in the application of identifying the mass terahertz spectroscopy.
[1] LIU Sheng-gang, ZHONG Ren-bin(刘盛纲, 钟任斌). Journal of University of Electronic Science and Technology of China(电子科技大学学报), 2009, 38(5): 481. [2] Tonouchi M. Nature Photonics, 2007, 1(2): 97. [3] Bax ter J B, Guglietta G W. Analytical Chemistry, 2011, 83(12): 4342. [4] Zhong H, Redo-Sanchez A, Zhang X C. Optics Express, 2006, 14(20): 9130. [5] Liang M, Shen J, Wang G. Journal of Physics D: Applied Physics, 2008, 41(13): 135306. [6] LIANG Mei-yan, ZHAO Shu-sen, SHEN Jing-ling(梁美彦, 赵树森, 沈京玲). Acta Optica Sinica(光学学报), 2009, 29(s1): 226. [7] Pan R, Zhao S, Shen J. Procedia Engineering, 2010, 7: 15. [8] Chen T, Li Z, Mo W. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2013, 106: 48. [9] Ge H, Jiang Y, Xu Z, et al. Optics Express, 2014, 22(10): 12533. [10] Bengio Y, Courville A, Vincent P. Pattern Analysis and Machine Intelligence, IEEE Transactions on, 2013, 35(8): 1798. [11] Hinton G. Momentum, 2010, 9(1): 926. [12] LIU Chuan-wu, ZHANG Zhi-jun, BI Du-yan(刘传武, 张智军, 毕笃彦). Systems Engineering and Electronics(系统工程与电子技术), 2009, 8: 16. [13] Guo G, Wang H, Bell D, et al. Computational Linguistics and Intelligent Text Processing. Springer Berlin Heidelberg, 2004. 559. [14] ZHANG Lei, LIU Jian-wei, LUO Xiong-lin(张 磊, 刘建伟, 罗雄麟). Pattern Recognition and Artificial Intelligence(模式识别与人工智能), 2010, 23(3): 376. [15] Vincent P, Larochelle H, Bengio Y, et al. Proceedings of the 25th International Conference on Machine Learning. ACM, 2008. 1096.
