大鼠胰腺及癌组织红外光谱连续小波特征提取及径向基人工神经网络识别

采用水平衰减全反射(HATR)傅里叶变换红外光谱法(FTIR)测定了SD大鼠胰腺正常组织与非正常组织的谱图,提出了一种新的基于FTIR的连续小波特征提取与径向基人工神经网络分类方法以提高FTIR对早期SD大鼠胰腺癌的诊断准确率。利用连续小波多分辨率分析法提取FTIR特征量,对于提取的特征量采用径向基函数神经网络进行模式分类。对SD大鼠的胰腺正常组织、早期癌组织及进展期癌组织的FTIR,利用连续小波多分辨率分析法提取9个特征量,进行RBF神经网络分类判断。当目标误差为0.01,径向基函数的分布常数为5时,网络达到最优化,总的正确识别率为96.67%。并对影响分类结果的网络参数、目标误差和分布常数对分类样品的影响做了讨论。实验结果表明:此方法对早期胰腺癌具有较高的诊断率。     

应用小波特征提取肺癌组织FTIR的支持向量机分类方法研究

提出了一种新的基于傅里叶变换红外光谱(Fourier Transform Infrared Spectroscopy, FTIR)的小波特征提取与支持向量机(SVM)分类方法以提高FTIR对早期肺癌的诊断准确率. 对肺正常组织、早期肺癌及进展期肺癌组织的FTIR, 利用连续小波(CW)多分辨率分析法提取9个特征量, 支持向量机把其分为正常组与非正常组(包括早期肺癌和进展期肺癌), 对正常组织、早期肺癌和进展期肺癌的识别, 多项式核函数和径向基函数的识别准确率最高. 多项式核函数对正常组织、早期肺癌和进展期肺癌的识别准确率分别为100%, 95%及100%; 径向基函数分别为100%, 95%和100%. 实验结果表明FTIR-CW-SVM模式分类方法对正常肺癌组织、早期肺癌及进展肺癌的识别具有较好的可行性.     

一种新颖的基于FTIR-CWT及ANN分类法的同科属中药材菟丝子与金灯藤的识别研究

采用水平衰减全反射傅里叶变换红外光谱法（HATR-FTIR）测定了同属种子植物中药材菟丝子及金灯藤的FTIR,运用基于连续小波多分辨率分析法对吸收较为相似的菟丝子及金灯藤的FTIR进行特征提取。选择第7、10、13分解层数的特征向量,进行人工神经网络（ANN）训练,再用训练出来的网络对不同产地的植物种子菟丝子和金灯藤所得FTIR小波提取的特征向量进行分类。通过对32个不同样本的验证,说明能够采用基于FTIR-连续小波特征提取及人工神经网络分类法对同科属中药材菟丝子与金灯藤进行识别。     

Back‐propagation artificial neural network and attenuated total reflectance‐Fourier transform infrared spectroscopy for diagnosis of basal cell carcinoma by blood sample analysis

A diagnostic method for the cancer, based on investigation of infrared spectra of blood samples, has been developed. The two‐layer modified principal component feed forward back‐propagation artificial neural network (BP‐ANN) was used to classify the attenuated total reflectance‐Fourier transform infrared (ATR‐FTIR) spectra of blood samples obtained from healthy people and those with basal cell carcinoma (BCC). Results showed 98.33% of accuracy, in comparison with the current clinical methods. In the first step, 20 blood samples (10 normal and 10 cancer cases) were applied to construct the calibration model. Spectroscopic studies were performed in 900–1800 cm⁻¹ spectral region with 3.85 cm⁻¹ data space. In order to modify the capability of ANN in prediction of test samples, two different algorithms were applied. The obtained results confirmed the compatibility of the proposed network with the architecture of 20‐8‐2 (input‐hidden‐output) with the pattern model. It was concluded that analysis of blood samples by ATR‐FTIR spectroscopy and ANN chemometric technique would be a reliable approach for detection of BCC. Copyright © 2009 John Wiley & Sons, Ltd.     

水平衰减全反射-傅里叶变换红外光谱结合化学计量学法应用于三种缩叶藓属植物的分类

采用水平衰减全反射(HATR)-傅里叶变换红外光谱法(FTIR)测定了3种缩叶藓属植物齿边缩叶藓、多枝缩叶藓和中华缩叶藓的红外谱图,运用离散小波变换对吸收较为相似的3种缩叶藓属植物的红外谱图进行特征提取。通过分析比较后选择第三,四分解层进行特征向量的提取,利用所得到的特征变量进行径向基神经网络(RBF-NN)训练,再将训练出来的网络对不同产地的3种缩叶藓属植物的红外谱图离散小波提取后的特征向量进行分类。通过对120个不同样本的验证,说明能够采用基于FTIR-离散小波进行数据压缩后进行特征变量的提取及径向基神经网络分类法对3种缩叶属植物齿边缩叶藓、多枝缩叶藓和中华缩叶藓进行分类。     

Identification of epilepsy from intracranial EEG signals by using different neural network models

In this work, a framework is provided for identifying intracranial electroencephalography (iEEG) seizures based on discrete wavelet transform (DWT) analysis of iEEG signals using forward propagation and feedback neural networks. The performance of 5 different data sets combination classifications is studied using the probabilistic neural network (PNN), learning vector quantization neural network (LVQ) and Elman neural network (ENN). Different feature combinations serve as the input vectors of the classifiers to obtain the best outcomes. It has been found that PNN has less running time and provides better classification accuracy (CA) than ENN and LVQ classifiers for all 5 classification problems. It is worth noticing that the CA for the C-D classification task, which shows the status of pre-ictal versus post-ictal, has been greatly improved, and reached 83.13%. Hence, the epilepsy iEEG signals pattern recognition based on DWT statistical features using the PNN classifier is more suitable for forming a reliable, automatic classification system in order to assist doctors in diagnosis.     

Detection of Cholangiocarcinoma with Fourier Transform Infrared Spectroscopy and Radial Basis Function Neural Network Classification

The aim of this study was to explore the possibility of applying Fourier transform infrared(FTIR) spectroscopy as a medical diagnostic tool based on a neural network classifier for detecting and classifying cholangiocarcinoma. A total of 51 cases of bile duct tissues were obtained and later characterized by FTIR spectroscopy prior to pathological diagnosis. The criteria for classification included 30 parameters for each FTIR spectra, including peak position(P), intensity(I) and full width at half-maximum(FWHM), were measured, calculated and subsequently compared against the normal and cancer groups. The FTIR spectra were classified by the radial basis function(RBF) network model. For establishing the RBF, 23 cases were used to train the RBF classifier, and 28 cases were applied to validate the model. Using the RFB model, nine parameters were observed to be pronouncedly different between cancerous and normal tissue, including I₁₆₄₀, I₁₅₅₀, I₁₄₆₀, I₁₄₀₀, I₁₂₅₀, I₁₁₂₀, I₁₀₈₀, I₁₀₄₀ and P₁₀₄₀. In the RBF training classification, the accuracy, sensitivity, and specificity of diagnosis were 82.6%, 80.0%, and 84.6%, respectively. While validating the classification, the accuracy, sensitivity, and specificity of diagnosis were 78.6%, 75.0%, and 81.2%, respectively. The results suggest that FTIR spectroscopy combined with neural network classifier could be applied as a medical diagnostic tool in cholangiocarcinoma diagnosis.     

Identification between Fimbristylis miliacea (L.) Vahl and Fimbristhlis stauntonii Debeaux et Franch.by CWT-FTIR-RBFNN

Horizontal attenuation total reflection‐Fourier transform infrared spectroscopy (HATR‐FTIR) is used to measure the FTIR of Fimbristylis miliacea (L.) Vahl seed and Fimbristhlis stauntonii Debeaux et Franch. seed. In order to extrude the difference between Fimbristylis miliacea (L.) Vahl seed and Fimbristhlis stauntonii Debeaux et Franch. seed, continuous wavelet transform (CWT) is used to decompose the FTIR of Fimbristylis miliacea (L.) Vahl seed and Fimbristhlis stauntonii Debeaux et Franch. seed. Three main scales are selected as the feature extracting space in the CWT domain. According to the distribution of FTIR of Fimbristylis miliacea (L.) Vahl seed and Fimbristhlis stauntonii Debeaux et Franch. seed, three feature regions are determined at every spectra band at selected three scales in the CWT domain. Thus nine feature parameters form the feature vector. The feature vector is input to the radial basis function neural network (RBFNN) to train so as to accurately classify the Fimbristylis miliacea (L.) Vahl seed and Fimbristhlis stauntonii Debeaux et Franch. seed. 110 couples of FTIR are used to train and test the proposed method, where 60 couples are used as training samples and 50 couples are used as testing samples. Experimental results show that the accurate recognition rate between Fimbristylis miliacea (L.) Vahl seed and Fimbristhlis stauntonii Debeaux et Franch. seed is respectively 96% and 98% by using the proposed method.     

人工神经网络在示波计时电位法中的应用

研究了误差反传神经网络和小波变换在示波计时电位测定中应用的可能性。对Ｐｂ２ ＋ 等无机离子测定的实验结果表明： 人工神经网络可较好地用于高次微分和经小波变换处理后的示波计时电位测定。 与经典的示波测定方法相比较, 该法基本上消除了人为误差的影响, 提高了分析速度和分析结果的可靠性。     

校正集样品特性对鼠药中毒鼠强近红外光谱分析的影响

根据市售鼠药样品成分各异且相对复杂,建立6种不同成分体系和9个不同样本容量的校正集,运用小波变换压缩鼠药的近红外透射光谱数据,结合BP反向神经网络算法对压缩的数据进行建模,考察校正集样品特性对模型预测能力的影响。试验结果表明:采用BP神经网络算法建立定量模型时,只要校正集样品中包含了与预测样品性质相似的样本,就能准确地对复杂样品进行近红外定量分析。当校正集容量分别为72和84时,模型预测结果趋于平稳。当校正集数量为96时,模型的最大相关系数为0.959 8,预测最小标准差和平均相对误差分别为1.893%和1.92%。     

Classification of Tieguanyin Tea with an Electronic Tongue and Pattern Recognition

Analysis of four Tieguanyin teas from different origins were performed using an electronic tongue, which has significant advantages in terms of accuracy and precision for pattern recognition. Hierarchical cluster analysis and principal component analysis were then applied to identify origins of these teas, and a distinct separation was observed. The back propagation neural network (BPNN) and the back propagation neural network with the Levenberg-Marquardt training algorithm (LMBP) were applied to build identification models. The Levenberg-Marquardt training algorithm model outperformed the back propagation neural network, as the identification performances of the former model were 100% in the training and prediction sets when four principal components were used. The results demonstrate that an electronic tongue with pattern recognition is suitable to classify Tieguanyin tea and shows broad potential in food inspection and quality control.     

离散小波特征提取及人工神经网络分类法的傅里叶变换红外光谱法识别鳞毛蕨科3种植物

利用水平衰减全反射-傅里叶变换红外光谱法测定了3种药用鳞毛蕨科植物贯众、阔鳞鳞毛蕨和变异鳞毛蕨根部的FT-IR。运用基于离散小波多分辨率分析法对FT-IR吸收较为相似的3种药用蕨类植物根的FT-IR进行特征提取。选择第4、5分解层数的特征向量,进行人工神经网络(Artificial neural network,ANN)训练;再用训练出来的网络对不同产地的3种药用蕨类植物根所得FT-IR小波提取的特征向量进行分类。通过对240个不同样本的预测,说明能够采用基于FT-IR-离散小波特征提取及人工神经网络分类法对同科3种药用蕨类植物根的FT-IR进行识别。     

Study on tribological behaviors of short glass fiber and polytetrafluoroethylene reinforced polycarbonate composites via a d‐optimal mixture design

This study analyzed variations of tribological behaviors that depend on the injection molding techniques during the blending of short glass fiber (SGF) and polytetrafluoroethylene (PTFE) reinforced polycarbonate (PC) composites. The proposed planning of blending experiments is to use a D‐optimal mixture design (DMD). The tribological behaviors of friction coefficient and wear mass loss were selected for discussion. Nine experimental runs, based on a DMD method, utilized to train the back‐propagation neural network (BPNN) and then the simulated annealing algorithm (SAA) approach is applied to search for an optimal mixture ratio setting. In addition, the result of BPNN integrating SAA was also compared with response surface methodology (RSM) approach. The results of confirmation experiment show that DMD, RSM, and BPNN integrating SAA method are effective tools for the optimization of reinforced process. Furthermore, the scanning electron microscope (SEM) images show that the abundant debris are peeled off from the matrix materials and predominant delamination mechanisms and plastic deformation are shown on the worn surface after tribological behavior tests. Copyright © 2010 John Wiley & Sons, Ltd.     

二进小波神经网络用于示波计时电位测定

提出了二进小波神经网络的结构及算法，并用于单组分和多组分示波计时电位信号的浓度计算。在二进小波神经网络中选用了Morlet母小波和修理的误差反传前向神经网络。探讨了二进小波神经网络中的中小波基个数，初始学习速率因子和动量因子等参数对网络预测结果的影响。结果表明：二进小波神经网络对双组分和单组分示波计时电位信号中去极剂浓度的预测均有很好效果。     

毛细管电泳-小波神经网络法辅佐诊断乳腺癌的研究

采用毛细管电泳法测定了46个健康人和26个乳腺癌病人尿样中的13种正常核苷和修饰核苷,以小波神经网络作为模式识别工具对健康人和乳腺癌病人的分类作了研究,随机选取的训练集的识别率达到100％,相应的预测集判别率正确性在96％以上,与经典的前向多层神经网络相比,小波神经网络具有更强的信息提取和逼近能力.研究结果还表明,小波神经网络的预测能力强于主成分分析和线性判别分析,毛细管电泳法与小波神经网络的结合有望成为乳腺癌的辅助诊断手段.     

基于小波神经网络的新型算法用于化学信号处理

基于紧支集正交小波神经网络的构造思想,用具有紧支集的Ｂ－样条函数的伸缩和平稳替代小波函数,提出了一种新型算法,并将其应用于化学信号的处理,实现了信号的压缩和滤噪,适应小波神经网络相比,学习速度得到了大幅度的提高。     

灰铸铁石墨形态的自动分类

在所提取的纹理特征的基础上,使用误差后向传播神经网络构建了一种优化的人工神经网络人顺。实现了灰铸铁石墨态的自动分类。用于描述石墨形态特征由分形维,粗细参数和二维自回归系数共同组成。该法成功地将人工神经网络引入了对灰铸铁石墨形态的分类,相对于传统人工目测法是一种很大的进步,而神经网络分类器的优化方法对其它神经网络模型的构也具有一定参考价值。     

Resolution of overlapping signals in spectrometry using a wavelet packet transform and an Elman recurrent neural network

A novel method named a wavelet packet transform based Elman recurrent neural network (WPTERNN) was proposed for the simultaneous UV–visible spectrometric determination of Cu(II), Cd(II) and Zn(II). This method combined wavelet packet denoising with an Elman recurrent neural network. A wavelet packet transform was applied to perform data compression, to extract relevant information, and to eliminate noise and collinearity. An Elman recurrent network was applied for nonlinear multivariate calibration. In this case, using trials, the kind of wavelet function, the decomposition level, and the number of hidden nodes for the WPTERNN method were selected as Daubechies 14, 3, and 8, respectively. A program (PWPTERNN) was designed that could perform the simultaneous determination of Cu(II), Cd(II) and Zn(II). The relative standard errors of prediction (RSEP) obtained for all components using WPTERNN, a Elman recurrent neural network (ERNN), partial least squares (PLS), principal component regression (PCR), Fourier transform based PCR (FTPCR), and multivariate linear regression (MLR) were compared. Experimental results demonstrated that the WPTERRN method was successful even where there was severe overlap of spectra. The results obtained from an additional test case also demonstrated that the WPTERNN method performed very well. Figure The part of WP coefficients obtained by wavelet packet transforms     

Classification of rapeseed and soybean oils by use of unsupervised pattern-recognition methods and neural networks

Unsupervised pattern-recognition methods and Kohonen neural networks have been applied to the classification of rapeseed and soybean oil samples according to their type and quality by use of chemical and physical properties (density, refractive index, saponification value, and iodine and acid numbers) and thermal properties (thermal decomposition temperatures) as variables. A multilayer feed-forward (MLF) neural network (NN) has been used to select the most important variables for accurate classification of edible oils. To accomplish this task different neural networks architectures trained by back propagation of error method, using chemical, physical, and thermal properties as inputs, were employed. The network with the best performance and the smallest root mean squared (RMS) error was chosen. The results of MLF network sensitivity analysis enabled the identification of key properties, which were again used as variables in principal components analysis (PCA), cluster analysis (CA), and in Kohonen self-organizing feature maps (SOFM) to prove their reliability.