Study On The Methods Of Contiuous Wavelet Transform In The FTIR

Spectral chemistry is an area which was very concerned by analysis chemistry and organic chemistry, of which Fourier transform infrared spectroscopy (FTIR) can get nearly all of material information about complex systems, so it has been widely applicated in various fields. How to make use of the large amount of data of absorption spectra from complex system for fast qualitative and quantitative analysis has been a goal of analytical chemists. Hence comes the chemical stoichiometry which was widely used in many fields because of the organic combination in the numerical computing technology and information from the equipment. Wavelet transform is a more effective signal processing method after the Fourier transform, the transformed results (wavelet factor) of continuous wavelet transform (CWT) contain more valuable information, which is a relatively effective analysis methods of chemical stoichiometry. In this paper, plants for traditional Chinese drugs and cancer tissues were analyzed, the method of Horizontal Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy was used to obtain FTIR of samples, then they were analyzed and studied by analysis methods of chemical stoichiometry.1. Two Sibling Species seed of Cuscuta Chinensis Lam. and Cuscuta japonica Choisy, with two different ferns Equisetum arvense L. and Hippochaete ramosissima (Desf.) Boerner were chosen as medicinal plant and determined directly by HART-FTIR, then analyzed and studied by analysis methods of chemical stoichiometry. It provides a basis for the possibility of the classification of medicinal fern by continuous wavelet feature extraction method which has been applied to FTIR. In addition, Cyrtomium fortunei J. Sm in six different times and Woodwardia japonica (L.f.) Sm. in five different periods were collected from Zhejiang Jinhua by system. Three Sibling Species of Polystichum tsus-simense (Hook.) J. Smith., Polystichum acanthophyllum (Franch.) Christ and Cyrtomium fortunei J. Sm were collected from Xi'an. Ten different Sibling Species of Pteris multifida Poir, Onychium japonicum (Thunb.) Kze, Dryopteris fuscipes C. Chr., Pteris plumbea Chris, Pteris vuttata L., Coniogramme japonica (Thunb.) Diels, Microsorum fortunei (T.Moore) Ching, Lygodium japonicum (Thunb.) Sw and Phegopteris decursive-pinnata Fee were collected from Jinhua city Zhejiang province, using direct determination of HART-FTIR method to get the infrared spectra of they different parts, by choosing eigenvalue of infrared spectroscopy, then adopting statistical software respectively to do the clustering and principal component analysis, and combining with continuous wavelet analysis for feature extraction to Fourier transform infrared spectroscopy, as based on support vector machine classification, the results show that it is consistent with the traditional classification. It provides more scientific research basis for a combination of spectroscopy with computer science and technology in identification of traditional Chinese medicine and plant taxonomy in a wider range of applications.2. Normal tissues of the human lung, early lung cancer and advanced lung cancer; Sprague-Dawley (SD) rat's normal, early pancreatic and abnormal pancreatic tissues; SD rat's normal, dysplasia, early carcinoma and advanced carcinoma tissues samples were chosen for application of early diagnosis of cancer, the samples were determined directly by the HART-FTIR and extracted using wavelet feature, then using the radial basis function artificial neural network (RBFNN) for identification and study with support vector machine (SVM) method for classification. Wavelet transform was used to enlarge the difference of different micro-FTIR map, so that we can identify the normal tissue, early carcinoma and advanced carcinoma by machine learning algorithm, and then to establish a rapid accurate method for early diagnosis of cancer.