Multivariate image analysis: A review with applications

Nowadays, image analysis is becoming more important because of its ability to perform fast and non-invasive low-cost analysis on products and processes. Image analysis is a wide denomination that encloses classical studies on gray scale or RGB images, analysis of images collected using few spectral channels (sometimes called multispectral images) or, most recently, data treatments to deal with hyperspectral images, where the spectral direction is exploited in its full extension. Pioneering data treatments in image analysis were applied to simple images mainly for defect detection, segmentation and classification by the Computer Science community. From the late 80s, the chemometric community joined this field introducing powerful tools for image analysis, which were already in use for the study of classical spectroscopic data sets and were appropriately modified to fit the particular characteristics of image structures. These chemometric approaches adapt to images of all kinds, from the simplest to the hyperspectral images, and have provided new insights on the spatial and spectroscopic information of this kind of data sets. New fields open by the introduction of chemometrics on image analysis are exploratory image analysis, multivariate statistical process control (monitoring), multivariate image regression or image resolution. This paper reviews the different techniques developed in image analysis and shows the evolution in the information provided by the different methodologies, which has been heavily pushed by the increasing complexity of the image measurements in the spatial and, particularly, in the spectral direction.     

Elemental and molecular characterization of aged polydimethylsiloxane foams

The application and integration of micro X-ray fluorescence (MXRF) and Fourier transform infrared (FT-IR) imaging to polydimethylsiloxane (PDMS) foam aging issues have been applied to cross-sectional images. Previous work has shown the tin in the stannous 2-ethylhexanoate catalyst to be highly mobile and it typically migrates to the PDMS foam upper surface. The current paper discusses a method for the integration of full spectral MXRF and FT-IR imaging of aged foams. Solvent extractions have also been performed on both fresh and aged foams to further examine aged foam properties. Combining elemental and molecular imaging techniques and applying them to PDMS aging provides synergistic information that aids in understanding the sample composition and distribution of components. Application of chemometric analysis to the full spectral elemental and molecular maps demonstrates correlations within the foams of the residual tin, organo-tin functional group moieties, and the presence of nitroplasticizer from an exogenous source.     

Development of 2D band-target entropy minimization and application to the deconvolution of multicomponent 2D nuclear magnetic resonance spectra

Spectral reconstruction from multicomponent spectroscopic data is the frequent primary goal in chemical system identification and exploratory chemometric studies. Various methods and techniques have been reported in the literature. However, few algorithms/methods have been devised for spectral recovery without the use of any a priori information. In the present studies, a higher dimensional entropy minimization method based on the BTEM algorithm (Widjaja, E.; Li, C.; Garland, M. Organometallics 2002, 21, 1991-1997.) and related techniques were extended to large-scale arrays, namely, 2D NMR spectroscopy. The performance of this novel method had been successfully verified on various real experimental mixture spectra from a series of randomized 2D NMR mixtures (COSY NMR and HSQC NMR). With the new algorithm and raw multicomponent NMR alone, it was possible to reconstruct the pure spectroscopic patterns and calculate the relative concentration of each species without recourse to any libraries or any other a priori information. The potential advantages of this novel algorithm and its implications for general chemical system identification of unknown mixtures are discussed.     

Chemical imaging of iron oxides and oxyhydroxides using near infrared Raman imaging microscopy

Abstract

A new method of identifying and mapping the distribution of both iron oxides and oxyhydroxides using near infrared Raman imaging microscopy NIRIM is reported. This technique offers an important alternative to conventional spectroscopic techniques that provide spatially averaged data. The NIRIM instrument used for these studies combines fibre bundle image compression hardware and multivariate signal processing software to identify and map different iron minerals and corrosion products. The NIRIM images clearly distinguish hematite, magnetite, wustite, goethite, and lepidocrocite microstructures. The first chemical maps of naturally occurring iron minerals and corroded steel surfaces obtained using NIRIM spectral images, classified with the aid of a library of pure compound spectra, are presented.     

2D correlation spectroscopy and multivariate curve resolution in analyzing pH-dependent evolving systems monitored by FT-IR spectroscopy,a comparative study

Multivariate curve resolution (MCR) and 2D correlation spectroscopy (2D-CoS), including sample-sample correlation, have been applied to the analysis of evolving midinfrared spectroscopic data sets obtained from titrations of organic acids in aqueous solution. In these data sets, well-defined species with significant differences in their spectra are responsible for the spectral variation observed. The two fundamentally different chemometric techniques have been evaluated and discussed on the basis of experimental and supportive simulated data sets. MCR gives information that can be directly related to the chemical species that is of importance from a practical point of view, whereas 2D-CoS results normally require more interpretation. The obtained conclusions are regarded valid for similar evolving data, which are increasingly being encountered in analytical chemistry when multivariate detectors are used to follow dynamic processes, including separations as well as chemical reactions, among others.     

Chemical imaging of latent fingerprint residues

In situ attenuated total reflection Fourier transform infrared (ATR-FT-IR) spectroscopic imaging has been used to obtain chemical images of fingerprints under controlled humidity and temperature. The distribution of lipid and amino acid components in the fingerprints from different donors left on the surface of the ZnSe crystal has been studied using an in situ FT-IR spectroscopic imaging approach under a controlled environment and studied as a function of time. Univariate and multivariate analyses were employed to analyze the spectroscopic dataset. Changes in the spectra of lipids with temperature and time have been detected. This information is needed to understand aging of the fingerprints. The ATR-FT-IR spectroscopic imaging offers a new and complementary means for studying the chemistry of fingerprints that are left pristine for further analysis. This study demonstrates the potential for visualizing the chemical changes of fingerprints for forensic applications by spectroscopic imaging.     

Evaluation of the orthogonal projection on latent structure model limitations caused by chemical shift variability and improved visualization of biomarker changes in 1H NMR spectroscopic metabonomic studies

In general, applications of metabonomics using biofluid NMR spectroscopic analysis for probing abnormal biochemical profiles in disease or due to toxicity have all relied on the use of chemometric techniques for sample classification. However, the well-known variability of some chemical shifts in 1H NMR spectra of biofluids due to environmental differences such as pH variation, when coupled with the large number of variables in such spectra, has led to the situation where it is necessary to reduce the size of the spectra or to attempt to align the shifting peaks, to get more robust and interpretable chemometric models. Here, a new approach that avoids this problem is demonstrated and shows that, moreover, inclusion of variable peak position data can be beneficial and can lead to useful biochemical information. The interpretation of chemometric models using combined back-scaled loading plots and variable weights demonstrates that this peak position variation can be handled successfully and also often provides additional information on the physicochemical variations in metabonomic data sets.     

MCRC software: A tool for chemometric analysis of two-way chromatographic data

This paper describes the development and implementation of MCRC software, chemometric software for Multivariate Curve Resolution of two-way Chromatographic data. MCRC software is developed for chemometric analysis of chromatographic data; however, it may also be used for other types of multivariate data. It consists of five groups of techniques for preprocessing, chemical rank determination, local rank analysis, multivariate resolution and peak integration. This software has the ability of the analysis of complex multi-component chromatographic signals of gas chromatography-mass spectrometry (GC-MS) and high performance liquid chromatography-diode array detection (HPLC-DAD). The software allows a user to apply the implemented methods in an easy way and it gives a straightforward possibility to visualize the obtained results. The main features of the presented software are providing a number of preprocessing techniques, implementation of different chemical rank determination methods, usage of iterative and non-iterative resolution techniques and a user-friendly environment with a variety of graphical outputs. The implementation of the MCRC software is demonstrated by the analysis of an overlapped peak cluster of simulated GC-MS data.     

Spectroscopic imaging of latent fingermarks collected with the aid of a gelatin tape

This work introduces a new and nondestructive methodology for the collection and chemical identification of latent fingermarks. The main challenges of this work were (a) to find an appropriate medium to lift fingermarks from various surfaces and (b) to develop an analytical approach for the identification of small quantities of sample while avoiding spectroscopic interference from the lifting media. Two different lifting media were evaluated and analyzed by ATR-FT-IR spectroscopic imaging, which affords inherent chemical specificity with rapid acquisition of data. This is the first time that chemical images of latent fingermarks collected with gel lifters from different surfaces have been obtained. Spatially resolved chemical images from different depths within the same sample were obtained using ATR-FT-IR imaging with a variable angle ATR accessory to minimize interference from the substrate. The possibility of obtaining, through the developed methodology, three-dimensional depth profiles of surface contaminants collected with the lifting gel shows great potential for the investigation of samples for forensic interest.     

Active DLP hyperspectral illumination: a noninvasive, in vivo, system characterization visualizing tissue oxygenation at near video rates

We report use of a novel hyperspectral imaging system utilizing digital light processing (DLP) technology to noninvasively visualize in vivo tissue oxygenation during surgical procedures. The system's novelty resides in its method of illuminating tissue with precisely predetermined continuous complex spectra. The Texas Instruments digital micromirror device, DMD, chip consisting of 768 by 1024 mirrors, each 16 μm square, can be switched between two positions at 12.5 kHz. Switching the appropriate mirrors controls the intensity of light illuminating the tissue as a function of wavelength, active spectral illumination. Meaning, the tissue can be illuminated with a different spectrum of light within 80 μs. Precisely, predetermined spectral illumination penetrates into patient tissue, its chemical composition augments the spectral properties of the light, and its reflected spectra are detected and digitized at each pixel detector of a silicon charge-coupled device, CCD. Using complex spectral illumination, digital signal processing and chemometric methods produce chemically relevant images at near video rates. Specific to this work, tissue is illuminated spectrally with light spanning the visible electromagnetic spectrum (380 to 780 nm). Spectrophotometric images are detected and processed visualizing the percentage of oxyhemoglobin at each pixel detector and presented continuously, in real time, at 3 images per second. As a proof of principle application, kidneys of four live anesthetized pigs were imaged before, during, and after renal vascular occlusion. DLP Hyperspectral Imaging with active spectral illumination detected a 64.73 ± 1.5% drop in the oxygenation of hemoglobin within 30 s of renal arterial occlusion. Producing chemically encoded images at near video rate, time-resolved hyperspectral imaging facilitates monitoring renal blood flow during animal surgery and holds considerable promise for doing the same during human surgical interventions.     

A spectral identity mapper for chemical image analysis

Generating chemically relevant image contrast from spectral image data requires multivariate processing algorithms that can categorize spectra according to shape. Conventional chemometric techniques like inverse least squares, classical least squares, multiple linear regression, principle component regression, and multivariate curve resolution are effective for predicting the chemical composition of samples having known constituents, but they are less effective when a priori information about the sample is unavailable. We have developed a multivariate technique called spectral identity mapping (SIM) that reduces the dependence of spectral image analysis on training datasets. The qualitative SIM method provides enhanced spectral shape specificity and improved chemical image contrast. We present SIM results of spectral image data acquired from polymer-coated paper substrates used in the manufacture of pressure sensitive adhesive tapes. In addition, we compare the SIM results to results from spectral angle mapping (SAM) and cosine correlation analysis (CCA), two closely related techniques.     

Fusion of MRI and CT images using guided image filter and image statistics

In medical imaging using different modalities such as MRI and CT, complementary information of a targeted organ will be captured. All the necessary information from these two modalities has to be integrated into a single image for better diagnosis and treatment of a patient. Image fusion is a process of combining useful or complementary information from multiple images into a single image. In this article, we present a new weighted average fusion algorithm to fuse MRI and CT images of a brain based on guided image filter and the image statistics. The proposed algorithm is as follows: detail layers are extracted from each source image by using guided image filter. Weights corresponding to each source image are calculated from the detail layers with help of image statistics. Then a weighted average fusion strategy is implemented to integrate source image information into a single image. Fusion performance is assessed both qualitatively and quantitatively. Proposed method is compared with the traditional and recent image fusion methods. Results showed that our algorithm yields superior performance.     

Quantitative spectroscopic analysis of heterogeneous mixtures: the correction of multiplicative effects caused by variations in physical properties of samples

Spectral measurements of complex heterogeneous types of mixture samples are often affected by significant multiplicative effects resulting from light scattering, due to physical variations (e.g., particle size and shape, sample packing, and sample surface, etc.) inherent within the individual samples. Therefore, the separation of the spectral contributions due to variations in chemical compositions from those caused by physical variations is crucial to accurate quantitative spectroscopic analysis of heterogeneous samples. In this work, an improved strategy has been proposed to estimate the multiplicative parameters accounting for multiplicative effects in each measured spectrum and, hence, mitigate the detrimental influence of multiplicative effects on the quantitative spectroscopic analysis of heterogeneous samples. The basic assumption of the proposed method is that light scattering due to physical variations has the same effects on the spectral contributions of each of the spectroscopically active chemical components in the same sample mixture. On the basis of this underlying assumption, the proposed method realizes the efficient estimation of the multiplicative parameters by solving a simple quadratic programming problem. The performance of the proposed method has been tested on two publicly available benchmark data sets (i.e., near-infrared total diffuse transmittance spectra of four-component suspension samples and near-infrared spectral data of meat samples) and compared with some empirical approaches designed for the same purpose. It was found that the proposed method provided appreciable improvement in quantitative spectroscopic analysis of heterogeneous mixture samples. The study indicates that accurate quantitative spectroscopic analysis of heterogeneous mixture samples can be achieved through the combination of spectroscopic techniques with smart modeling methodology.     

Restoration and spectral recovery of mid-infrared chemical images

Fourier transform infrared (FTIR) microspectroscopy is a powerful technique for label-free chemical imaging that has supplied important chemical information about heterogeneous samples for many problems across a variety of disciplines. State-of-the-art synchrotron based infrared (IR) microspectrometers can yield high-resolution images, but are truly diffraction limited for only a small spectral range. Furthermore, a fundamental trade-off exists between the number of pixels, acquisition time and the signal-to-noise ratio, limiting the applicability of the technique. The recently commissioned infrared synchrotron beamline, infrared environmental imaging (IRENI), overcomes this trade off and delivers 4096-pixel diffraction limited IR images with high signal-to-noise ratio in under a minute. The spatial oversampling for all mid-IR wavelengths makes the IRENI data ideal for spatial image restoration techniques. Here, we measured and fitted wavelength-dependent point-spread-functions (PSFs) at IRENI for a 74× objective between the sample plane and detector. Noise-free wavelength-dependent theoretical PSFs are deconvoluted from images generated from narrow bandwidths (4 cm(-1)) over the entire mid-infrared range (4000-900 cm(-1)). The stack of restored images is used to reconstruct the spectra. Restored images of metallic test samples with features that are 2.5 μm and smaller are clearly improved in comparison to the raw data images for frequencies above 2000 cm(-1). Importantly, these spatial image restoration methods also work for samples with vibrational bands in the recorded mid-IR fingerprint region (900-1800 cm(-1)). Improved signal-to-noise spectra are reconstructed from the restored images as demonstrated for a mixture of spherical polystyrene beads in a polyurethane matrix. Finally, a freshly thawed retina tissue section is used to demonstrate the success of deconvolution achievable with a heterogeneous, irregularly shaped, biologically relevant sample with distinguishing spectroscopic features across the entire mid-IR spectral range.     

Spectroscopic imaging of laser-induced plasma

Spectroscopic imaging provides 2D images with full spectral resolution at each pixel. Thus, chemical imaging of an object, as well as other useful information, can be obtained. An imaging spectroscopy method in the visible range is presented and applied to laser plasma. This is a powerful research tool with numerous possible applications. This study is focused on spectroscopic imaging of laser-produced plasmas, and such spectral images (full spectrum at each pixel) are presented for the first time. Detailed information on optical and geometrical effects are obtained, and an insight to the optimization of the laser plasma spectroscopy method is achieved. The size and the spatial shape of the plasma, which can be used for matrix effect compensation, are measured. Similarity maps and classification maps of laser-induced plasma are obtained for the first time. These maps are used for allocation of chemical components in the plasma. The signal to noise ratio maps of the spectra obtained from laser-induced plasmas are provided. These surfaces possess a clear maximum, indicating that there is a preferred site in the plasma, where the emitted light provides the best signal to noise ratio. The performance of the current method is limited by the lack of temporal resolution, although it can be extended by a proper temporal gating.     

Adaptable multivariate calibration models for spectral applications

Multivariate calibration techniques have been used in a wide variety of spectroscopic situations. In many of these situations, spectral variation can be partitioned into separate classes. For example, suppose that multiple spectra are obtained from each of a number of different objects wherein the level of the analyte of interest varies within each object over time. In such situations, the total spectral variation observed across all measurements has two distinct general sources of variation: intraobject and interobject. One might want to develop a global multivariate calibration model that predicts the analyte of interest accurately both within and across objects, including new objects not involved in developing the calibration model. However, this goal might be hard to realize if the interobject spectral variation is complex and difficult to model. If the intraobject spectral variation is consistent across objects, an effective alternative approach might be to develop a generic intraobject model that can be adapted to each object separately. This paper contains recommendations for experimental protocols and data analysis in such situations. The approach is illustrated with an example involving the noninvasive measurement of glucose using near-infrared reflectance spectroscopy. Extensions to calibration maintenance and calibration transfer are discussed.     

Fourier transform infrared (FT-IR) spectroscopy and improved principal component regression (PCR) for quantification of solid analytes in microalgae and bacteria

In bioanalytical chemistry, a detailed chemical understanding of biomaterials is often difficult to obtain due to the sheer number of analytes contained in a sample along with the samples' generally low reproducibility. This study presents a Fourier transform infrared (FT-IR) spectroscopic technique in conjunction with innovations in sample preparation and chemometric data preprocessing to overcome these limitations. These methodologies were applied to quantitative analyses of 31 representative compounds commonly found in biomaterial, which have been incorporated into a spectroscopic calibration database, that is, albumin (protein); D-alanine, glycine, histidine, valine, arginine, cysteine, phenylalanine, tyrosine, methionine, L-glutamine, and glutamic acid, (amino acids); glucose, fructose, galactose, mannose, sucrose, lactose, glycogen, agarose, and starch (carbohydrates); DNA (salmon sperm), sulphonoquinovosyl diglyceride ( sulpho-lipid ), and 1,2-diacyl-sn-glycero-3-phospho-L-serine ( phospho-lipid ); succinic acid and malic acid ( carboxylic acids ); glycolic acid (a -hydroxy acid), sodium pyruvate, b -carotene, frustules (microalgae silica-shells), and ammonium formate. Two proof-of-principle applications were based on calibration models incorporating these solids, i.e., characterization of E. coli and microalgae. The former aims for detection of bacterial contamination and the latter to enable investigations of changes in chemical composition of microalgae cells in response to shifting environmental conditions. Chemometric preprocessing steps have been developed for handling sample-to-sample fluctuations of absorption path lengths and baselines; the former incorporated mass normalization while the latter utilized a novel baseline correction method that requires no a priori information. Data preprocessing, chemometric calibration, and evaluation algorithms have been combined, together with an extensive spectral database of the aforementioned compounds (～1500 samples), for quantitative calibration purposes through the remotely accessible Virtual Chemometrics Lab , which can be utilized for a multitude of applications through a graphical user interface.     

A chemometric approach to detection and characterization of multiple protein conformers in solution using electrospray ionization mass spectrometry

Protein ion charge state distributions in electrospray ionization mass spectra have a potential to provide a wealth of information on protein dynamics, because they contain contributions from all protein conformers present in solution. Such ionic contributions often overlap, limiting the amount of useful information that can be extracted from the spectra. This difficulty is overcome in the present work by using a chemometric approach, which allows spectral deconvolution to be carried out and information on individual protein conformers to be extracted. Experiments are carried out by acquiring a series of spectra over a range of near-native and denaturing conditions to ensure significant changes in the conformers' populations. A total number of protein conformers sampled in the course of the experiments is determined by subjecting the set of collected spectra to singular value decomposition. Ionic contributions of each conformer to the total signal are then determined using a supervised optimization routine. Validation of the method has been carried out by monitoring acid- and alcohol-induced equilibrium states of well-characterized model proteins, chymotrypsin inhibitor 2 (two states), ubiquitin (three states) and apo-myoglobin (four states). For each of the model proteins, a new chemometric procedure yielded a picture of protein dynamics that was in excellent agreement with their documented behavior (as studied with other biophysical tools). The new method appears to be well-suited for monitoring protein dynamics in highly heterogeneous systems consisting of multiple proteins sampling a range of conformational states.     

Fast line-shape correction procedure for imaging Fourier-transform spectrometers

An instrument line-shape correction method adapted to imaging Fourier-transform spectrometers is demonstrated. The method calibrates all pixels on the same spectral grid and permits a direct comparison of the spectral features between pixels such as emission or absorption lines. Computation speed is gained by using matrix line-shape integration formalism rather than properly inverting the line shape of each pixel. A monochromatic source is used to characterize the spectral shift of each pixel, and a line-shape correction scheme is then applied on measured interferograms. This work is motivated by the emergence of affordable infrared CCD cameras that are currently being integrated in imaging Fourier-transform spectrometers.     

Simultaneous determination of the micro-, meso-, and macropore size fractions of porous polymers by a combined use of Fourier transform near-infrared diffuse reflection spectroscopy and multivariate techniques

Fourier transform near-infrared (FT-NIR) diffuse reflection spectroscopy was used in combination with principal component analysis and partial least-squares regression to simultaneously determine the physical and the chemical parameters of a porous poly(p-methylstyrene-co-1,2-bis(p-vinylphenyl)ethane) (MS/BVPE) monolithic polymer. Chemical variations during the synthesis of the polymer material can alter the pore volume and pore area distributions within the polymer scaffold. Furthermore, mid-infrared and near-infrared (NIR) spectroscopic chemical imaging was implemented as a tool to assess the uniformity of the samples. The presented study summarizes the comparative results derived from the spectral FT-NIR data combined with chemometric techniques. The relevance of the interrelation of physical and chemical parameters is highlighted whereas the amount of MS/BVPE (%, v/v) and the quantity (%) of micropores (diameter, d < 6 nm), mesopores (6 nm < d < 50 nm), and macropores (50 nm < d < 200 nm) could be determined with one measurement. For comparison of the quantitative data, the standard error of prediction (SEP) was used. The SEP for determining the MS/BVPE amount in the samples showed 0.35%, for pore volume quantiles 1.42-8.44%, and for pore area quantiles 0.38-1.45%, respectively. The implication of these results is that FT-NIR spectroscopy is a suitable technique for the screening of samples with varying physicochemical properties and to quantitatively determine the parameters simultaneously within a few seconds.