Maize kernel hardness classification by near infrared (NIR) hyperspectral imaging and multivariate data analysis

The use of near infrared (NIR) hyperspectral imaging and hyperspectral image analysis for distinguishing between hard, intermediate and soft maize kernels from inbred lines was evaluated. NIR hyperspectral images of two sets (12 and 24 kernels) of whole maize kernels were acquired using a Spectral Dimensions MatrixNIR camera with a spectral range of 960-1662 nm and a sisuChema SWIR (short wave infrared) hyperspectral pushbroom imaging system with a spectral range of 1000-2498 nm. Exploratory principal component analysis (PCA) was used on absorbance images to remove background, bad pixels and shading. On the cleaned images, PCA could be used effectively to find histological classes including glassy (hard) and floury (soft) endosperm. PCA illustrated a distinct difference between glassy and floury endosperm along principal component (PC) three on the MatrixNIR and PC two on the sisuChema with two distinguishable clusters. Subsequently partial least squares discriminant analysis (PLS-DA) was applied to build a classification model. The PLS-DA model from the MatrixNIR image (12 kernels) resulted in root mean square error of prediction (RMSEP) value of 0.18. This was repeated on the MatrixNIR image of the 24 kernels which resulted in RMSEP of 0.18. The sisuChema image yielded RMSEP value of 0.29. The reproducible results obtained with the different data sets indicate that the method proposed in this paper has a real potential for future classification uses.     

Variable-weighted least-squares support vector machine for multivariate spectral analysis

Multivariate spectral analysis has been widely applied in chemistry and other fields. Spectral data consisting of measurements at hundreds and even thousands of analytical channels can now be obtained in a few seconds. It is widely accepted that before a multivariate regression model is built, a well-performed variable selection can be helpful to improve the predictive ability of the model. In this paper, the concept of traditional wavelength variable selection has been extended and the idea of variable weighting is incorporated into least-squares support vector machine (LS-SVM). A recently proposed global optimization method, particle swarm optimization (PSO) algorithm is used to search for the weights of variables and the hyper-parameters involved in LS-SVM optimizing the training of a calibration set and the prediction of an independent validation set. All the computation process of this method is automatic. Two real data sets are investigated and the results are compared those of PLS, uninformative variable elimination-PLS (UVE-PLS) and LS-SVM models to demonstrate the advantages of the proposed method.     

基于高光谱和化学计量学方法的除草剂残留分类识别研究

本研究基于高光谱技术和化学计量学方法,对薄荷叶上的异丙甲草胺、烟嘧(莠去津、敌草胺和砜嘧(精喹4类除草剂残留进行种类判别.高光谱成像光谱范围为450nm～950nm的可见-近红外区域.为降低噪音对光谱数据的干扰、提升判别精度,采用SG平滑和多元散射校正对高光谱曲线进行处理.利用主成分分析算法(PCA)对原始数据进行降维...     

Combination of FTIR spectral imaging and chemometrics for tumour detection from paraffin-embedded biopsies

FTIR spectral imaging was applied on formalin-fixed paraffin-embedded biopsies from colon and skin cancerous lesions. These samples were deposited onto different substrates (zinc selenide and calcium fluoride respectively) and embedded using two types of paraffin. Formalin fixation followed by paraffin embedding is the gold standard in tissue storage. It can preserve molecular structures and it is compatible with immunohistochemistry. However, paraffin absorption bands are significant in the mid-infrared region and can mask some molecular vibrations of the tissue. Direct data processing was applied on spectral images without any chemical dewaxing of the tissues. Extended Multiplicative Signal Correction was used to correct the spectral contribution from paraffin. For this purpose, the signal of paraffin was modelled using Principal Component Analysis and paraffin spectra were removed from the raw images based on an outlier detection. Then, pseudo-colour images were computed by K-means clustering in order to highlight histological structures of interest. This robust chemometrics methodology was applied on the two samples. Tumour areas were successfully demarcated from the rest of the tissue in both colon and skin independently of the embedding material and of the substrate.     

化学计量学方法在光谱解析中的应用进展

本文对近20年来化学计量学在光谱分析测试中的应用进行了评述。指出了化学计量学在光谱分析中的重要性和应用前景。引用文献共129篇。     

Prediction of some quality attributes of lamb meat using near-infrared hyperspectral imaging and multivariate analysis

The goal of this study was to explore the potential of near-infrared (NIR) hyperspectral imaging in combination with multivariate analysis for the prediction of some quality attributes of lamb meat. In this study, samples from three different muscles (semitendinosus (ST), semimembranosus (SM), longissimus dorsi (LD)) originated from Texel, Suffolk, Scottish Blackface and Charollais breeds were collected and used for image acquisition and quality measurements. Hyperspectral images were acquired using a pushbroom NIR hyperspectral imaging system in the spectral range of 900–1700 nm. A partial least-squares (PLS) regression, as a multivariate calibration method, was used to correlate the NIR reflectance spectra with quality values of the tested muscles. The models performed well for predicting pH, colour and drip loss with the coefficient of determination (R²) of 0.65, 0.91 and 0.77, respectively. Image processing algorithm was also developed to transfer the predictive model in every pixel to generate prediction maps that visualize the spatial distribution of quality parameter in the imaged lamb samples. In addition, textural analysis based on gray level co-occurrence matrix (GLCM) was also conducted to determine the correlation between textural features and drip loss. The results clearly indicated that NIR hyperspectral imaging technique has the potential as a fast and non-invasive method for predicting quality attributes of lamb meat.     

Detection of pathological aortic tissues by infrared multispectral imaging and chemometrics

Processing of multispectral images is becoming an important issue, especially in terms of data mining for disease diagnosis. We report here an original image analysis procedure developed in order to compare 42 infrared multispectral images acquired on human ascending aortic healthy and pathological tissues. Each image contained about 2500 infrared absorption spectra, each composed of 1641 variables (wavenumbers). To process this large data set, we have restricted the spectral window used to the 1800-950 cm(-1) spectral range and selected 100 spectra from the aortic media, which is the most altered part of the aortic tissue in aneurysms. Prior to this selection, a spectral quality test was performed to eliminate 'bad' spectra. Our data set was first subjected to a discriminant analysis, which allowed separation of aortic tissues in two groups corresponding respectively to normal and aneurysmal states. Then a K-means analysis, based on 20 groups, allowed reconstruction of infrared images using false-colours and discriminated between pathological and healthy tissues. These results demonstrate the usefulness of such data processing methods for the analysis and comparison of a set of spectral images.     

Erythromycin-excipients compatibility studies using the thermal analysis and dynamic thermal infrared spectroscopy coupled with chemometrics

Journal of Thermal Analysis and Calorimetry - This study aims to characterize erythromycin (ERY) estolate by thermogravimetry analysis and differential scanning calorimetry. For such a purpose,...     

Potential applications of functional data analysis in chemometrics

In spectroscopy the measured spectra are typically plotted as a function of the wavelength (or wavenumber), but analysed with multivariate data analysis techniques (multiple linear regression (MLR), principal components regression (PCR), partial least squares (PLS)) which consider the spectrum as a set of m different variables. From a physical point of view it could be more informative to describe the spectrum as a function rather than as a set of points, hereby taking into account the physical background of the spectrum, being a sum of absorption peaks for the different chemical components, where the absorbance at two wavelengths close to each other is highly correlated. In a first part of this contribution, a motivating example for this functional approach is given. In a second part, the potential of functional data analysis is discussed in the field of chemometrics and compared to the ubiquitous PLS regression technique using two practical data sets. It is shown that for spectral data, the use of B-splines proves to be an appealing basis to accurately describe the data. By applying both functional data analysis and PLS on the data sets the predictive ability of functional data analysis is found to be comparable to that of PLS. Moreover, many chemometric datasets have some specific structure (e.g. replicate measurements, on the same object or objects that are grouped), but the structure is often removed before analysis (e.g. by averaging the replicates). In the second part of this contribution, we suggest a method to adapt traditional analysis of variance (ANOVA) methods to datasets with spectroscopic data. In particular, the possibilities to explore and interpret sources of variation, such as variations in sample and ambient temperature, are examined. Copyright © 2008 John Wiley & Sons, Ltd.     

Correlation analysis of characteristic infrared spectral data of hydantoin derivatives: evidence for vibrational coupling

The characteristic vibrations (ν_CO and ν_CC) of a large number of hydantoin derivatives are reported. Especially the very fine correlations ν_CO(sym) versus ν_CO(asym) (r²=0.985) but also successful correlations of the vibration wave numbers to HAMMETT's substituent constants and some other experimental parameters (pK_s, OxPot, RedPot) as well, corroborate reassignments of previously obtained results [Monatsh. Chem. 92 (1961) 361] and prove the doublet obtained in the region of the CO stretching vibrations to be the symmetrical and anti-symmetrical vibrational modes of a mechanically coupled system of two quasi-symmetrical CO bonds.     

An advanced multivariate approach for processing X-ray fluorescence spectral and hyperspectral data from non-invasive in situ analyses on painted surfaces

In the last decades, in situ non-invasive analytical techniques have been widely used for the analysis of paintings. These techniques are useful to extensively map the surface in a non-invasive way, in order to identify the most representative areas to be sampled. When spectroscopic investigations, such as X ray fluorescence (XRF), are conducted, they usually imply the acquisition of a huge amount of measurements. Subsequently, all these data should be processed in situ, in order to immediately support the sampling strategies. To this aim, an appropriate and fast strategy for multivariate treatment of XRF spectral and hyperspectral data sets is presented, able to account for inter-correlation among variables, which is an issue of high importance for elemental analyses. The main advantage of the approach is that XRF spectral profiles are analysed directly, without computation of derived parameters, by means of principal component analysis (PCA). This procedure allows a fast interpretation of results that can be accomplished in situ. Particular attention was paid to the selection of proper spectral pre-treatments to be applied on data together with the use of several chemometric tools (peak alignment, spectra normalisation and exploratory analysis) aimed at improving the interpretation of XRF results. In addition, the application of multivariate exploratory analysis on XRF hyperspectral maps was studied by using an interactive brushing procedure. The multivariate approach was validated on data obtained from the analysis of the famous Renaissance panel painting “The Ideal City”, exhibited in Palazzo Ducale of Urbino, Italy.     

Fourier transform infrared imaging analysis in discrimination studies of squamous cell carcinoma

Oral squamous cell carcinoma (OSCC) of the oral cavity and oropharynx represents more than 95% of all malignant neoplasms in the oral cavity. Histomorphological evaluation of this cancer type is invasive and remains a time consuming and subjective technique. Therefore, novel approaches for histological recognition are necessary to identify malignancy at an early stage. Fourier transform infrared (FTIR) imaging has become an essential tool for the detection and characterization of the molecular components of biological processes, such as those responsible for the dynamic properties of tumor progression. FTIR imaging is a modern analytical technique enabling molecular imaging of a complex biological sample and is based on the absorption of IR radiation by vibrational transitions in covalent bonds. One major advantage of this technique is the acquisition of local molecular expression profiles, while maintaining the topographic integrity of the tissue and avoiding time-consuming extraction, purification, and separation steps. With this imaging technique, it is possible to obtain unique images of the spatial distribution of proteins, lipids, carbohydrates, cholesterols, nucleic acids, phospholipids, and small molecules with high spatial resolution. Analysis and visualization of FTIR imaging datasets are challenging and the use of chemometric tools is crucial in order to take advantage of the full measurement. Therefore, methodologies for this task based on the novel developed algorithm for multivariate image analysis (MIA) are often necessary. In the present study, FTIR imaging and data analysis methods were combined to optimize the tissue measurement mode after deparaffinization and subsequent data evaluation (univariate analysis and MIAs). We demonstrate that it is possible to collect excellent IR spectra from formalin-fixed paraffin-embedded (FFPE) tissue microarrays (TMAs) of OSCC tissue sections employing an optimised analytical protocol. The correlation of FTIR imaging to the morphological tissue features obtained by histological staining of the sections demonstrated that many histomorphological tissue patterns can be visualized in the colour images. The different algorithms used for MIAs of FTIR imaging data dramatically increased the information content of the IR images from squamous cell tissue sections. These findings indicate that intra-operative and surgical specimens of squamous cell carcinoma tissue can be characterized by FTIR imaging.     

Detection of illicit substances in fingerprints by infrared spectral imaging

FTIR and Raman spectral imaging can be used to simultaneously image a latent fingerprint and detect exogenous substances deposited within it. These substances might include drugs of abuse or traces of explosives or gunshot residue. In this work, spectral searching algorithms were tested for their efficacy in finding targeted substances deposited within fingerprints. “Reverse” library searching, where a large number of possibly poor-quality spectra from a spectral image are searched against a small number of high-quality reference spectra, poses problems for common search algorithms as they are usually implemented. Out of a range of algorithms which included conventional Euclidean distance searching, the spectral angle mapper (SAM) and correlation algorithms gave the best results when used with second-derivative image and reference spectra. All methods tested gave poorer performances with first derivative and undifferentiated spectra. In a search against a caffeine reference, the SAM and correlation methods were able to correctly rank a set of 40 confirmed but poor-quality caffeine spectra at the top of a dataset which also contained 4,096 spectra from an image of an uncontaminated latent fingerprint. These methods also successfully and individually detected aspirin, diazepam and caffeine that had been deposited together in another fingerprint, and they did not indicate any of these substances as a match in a search for another substance which was known not to be present. The SAM was used to successfully locate explosive components in fingerprints deposited on silicon windows. The potential of other spectral searching algorithms used in the field of remote sensing is considered, and the applicability of the methods tested in this work to other modes of spectral imaging is discussed.     

Characterisation of non-viable whole barley, wheat and sorghum grains using near-infrared hyperspectral data and chemometrics

Undesired germination of cereal grains diminishes process utility and economic return. Pre-germination, the term used to describe untimely germination, leads to reduced viability of a grain sample. Accurate and rapid identification of non-viable grain is necessary to reduce losses associated with pre-germination. Viability of barley, wheat and sorghum grains was investigated with near-infrared hyperspectral imaging. Principal component analyses applied to cleaned hyperspectral images were able to differentiate between viable and non-viable classes in principal component (PC) five for barley and sorghum and in PC6 for wheat. An OH stretching and deformation combination mode (1,920–1,940 nm) featured in the loading line plots of these PCs; this water-based vibrational mode was a major contributor to the viable/non-viable differentiation. Viable and non-viable classes for partial least squares-discriminant analysis (PLS-DA) were assigned from PC scores that correlated with incubation time. The PLS-DA predictions of the viable proportion correlated well with the viable proportion observed using the tetrazolium test. Partial least squares regression analysis could not be used as a source of contrast in the hyperspectral images due to sampling issues.     

Classification of spectroscopically encoded resins by Raman mapping and infrared hyperspectral imaging

Barcoded resins (BCRs) were recently introduced as a potential platform for pre-encoded multiplexed synthesis, screening, and biomedical diagnostics. A key step toward the development of this strategy is the ability to rapidly interrogate and classify the BCRs in a high-throughput, noninvasive manner. Here, we describe a one-step strategy based on Raman mapping and Fourier transform infrared imaging to classify and spatially resolve randomly distributed BCRs. To illustrate this methodology, mixtures of up to 25 different BCRs were imaged and classified with 100% confidence. This strategy can be readily extended to a larger pool of resins, provided each BCR features a unique vibrational fingerprint (spectroscopic barcode). We have also established that reliable single-bead Raman spectra can be recorded in 10 ms, thus confirming that Raman mapping, in particular, could be a very fast method to classify the BCRs.     

Laser Raman and infrared spectral studies of dl-phenylalaninium nitrate

The vibrational band assignments of dl-phenylalaninium nitrate in the crystalline state are made by recording the infrared and Raman spectra at room temperature. The presence of carbonyl (C=O) group has been identified. The prominent marker bands of the aromatic amino acid phenylalanine have been observed and the various modes of vibration have been assigned. The extensive intermolecular hydrogen bonding in the crystal has been identified by the shifting of bands due to the stretching and bending modes of the various functional groups. The nitrate group forms the anion. The stretching and bending wave numbers of the NO(3)(-) anion are different from those observed for free ion state and the degenerating mode of vibrations is also lifted. These reveal that the crystalline field has influenced the symmetry of the nitrate ion.     

Chalcones XVI: infrared spectral studies of phenanthryl chalcones

IR absorption spectra of ten phenanthryl chalcones have been studied with a view to see the effect on stretching frequencies of , -unsaturated carbonyl group when conjugated to the phenanthryl nucleus.     

Orthogonal projection to latent structures solution properties for chemometrics and systems biology data

Partial least squares (PLS) is a widely used algorithm in the field of chemometrics. In calibration studies, a PLS variant called orthogonal projection to latent structures (O‐PLS) has been shown to successfully reduce the number of model components while maintaining good prediction accuracy, although no theoretical analysis exists demonstrating its applicability in this context. Using a discrete formulation of the linear mixture model known as Beer's law, we explicitly analyze O‐PLS solution properties for calibration data. We find that, in the absence of noise and for large n, O‐PLS solutions are simpler but just as accurate as PLS solutions for systems in which analyte and background concentrations are uncorrelated. However, the same is not true for the most general chemometric data in which correlations between the analyte and background concentrations are nonzero and pure profiles overlap. On the contrary, forcing the removal of orthogonal components may actually degrade interpretability of the model. This situation can also arise when the data are noisy and n is small, because O‐PLS may identify and model the noise as orthogonal when it is statistically uncorrelated with the analytes. For the types of data arising from systems biology studies, in which the number of response variables may be much greater than the number of observations, we show that O‐PLS is unlikely to discover orthogonal variation whether or not it exists. In this case, O‐PLS and PLS solutions are the same. Copyright © 2011 John Wiley & Sons, Ltd.     

Supervised principal components: a new method for multivariate spectral analysis

The supervised principal components (SPC) method was proposed by Bair and Tibshirani for statistics regression problems where the number of variables greatly exceeds the number of samples. This case is extremely common in multivariate spectral analysis. The objective of this research is to apply SPC to near‐infrared and Raman spectral calibration. SPC is similar to traditional principal components analysis except that it selects the most significant part of wavelength from the high‐dimensional spectral data, which can reduce the risk of overfitting and the effect of collinearity in modeling according to a semi‐supervised strategy. In this study, four conventional regression methods, including principal component regression, partial least squares regression, ridge regression, and support vector regression, were compared with SPC. Three evaluation criteria, coefficient of determination (R²), external correlation coefficient (Q²), and root mean square error of prediction, were calculated to evaluate the performance of each algorithm on both near‐infrared and Raman datasets. The comparison results illustrated that the SPC model had a desirable ability of regression and prediction. We believe that this method might be an alternative method for multivariate spectral analysis. Copyright © 2013 John Wiley & Sons, Ltd.