Spectral Vectors Research Articles

QUANTIFICATION OF BINARY MIXTURES OF COCAINE AND ADULTERANTS USING DISPERSIVE RAMAN AND FT-IR SPECTROSCOPY AND PRINCIPAL COMPONENT REGRESSION

Current forensic methods for detecting and identifying cocaine and its adulterants are destructive, so evidence cannot be re-analyzed. Vibrational techniques [Raman and Fourier transform infrared (FT-IR) spectroscopy] allow rapid, economical, nondestructive analysis. This work compares these techniques as methods for quantifying mixtures of cocaine (in the crack presentation) and adulterants. The aim is to provide a method to determine the amount of cocaine in crack rocks and to identify and quantify possible adulterants. A sample of crack was adulterated with benzocaine, caffeine, sodium carbonate, and lidocaine to create binary mixtures of concentrations of 20%, 40%, 60%, and 80%; pure samples of each adulterant and of crack were also examined. All samples were observed using dispersive Raman and attenuated total reflectance FT-IR spectroscopy. Quantitative analysis was performed based on principal component regression (PCR) applied to simulated and real spectra. The PCR model revealed that the Raman spectra yielded only minor measurement errors and the highest correlations using the PC2 spectral vector, which presented spectral features of cocaine. Among the mixtures, the best and worst results were obtained for caffeine and sodium carbonate, respectively. Dispersive Raman spectroscopy outperformed FT-IR in the quantitative determination of binary mixtures of cocaine and adulterants.

GROUND OBJECT RECOGNITION USING COMBINED HIGH RESOLUTION AIRBORNE IMAGES AND DSM

Abstract. The research is carried on dataset Vaihingen acquired from ISPRS Test Project on Urban Classification and 3D Building Reconstruction. Four different types of ground objects are extracted: buildings, trees, vegetation (grass and low bushes) and road. Spectral information is used to classify the images and then a refinement process is carried out using DSM. A novel method called Sparse Representation is introduced to extract ground objects from airborne images. For each pixel we extract its spectral vector and solve Basis Pursuit problem using l1 minimization. The classification of the pixel is same as the column vector of observation matrix corresponding to the largest positive component of the solution vector. A refinement procedure based on elevation histogram is carried on to improve the coarse classification due to misclassification of trees/vegetation and buildings/road.

Evaluating Classification Techniques for Mapping Vertical Geology Using Field-Based Hyperspectral Sensors

Hyperspectral data acquired from field-based platforms present new challenges for their analysis, particularly for complex vertical surfaces exposed to large changes in the geometry and intensity of illumination. The use of hyperspectral data to map rock types on a vertical mine face is demonstrated, with a view to providing real-time information for automated mining applications. The performance of two classification techniques, namely, spectral angle mapper (SAM) and support vector machines (SVMs), is compared rigorously using a spectral library acquired under various conditions of illumination. SAM and SVM are then applied to a mine face, and results are compared with geological boundaries mapped in the field. Effects of changing conditions of illumination, including shadow, were investigated by applying SAM and SVM to imagery acquired at different times of the day. As expected, classification of the spectral libraries showed that, on average, SVM gave superior results for SAM, although SAM performed better where spectra were acquired under conditions of shadow. In contrast, when applied to hypserspectral imagery of a mine face, SVM did not perform as well as SAM. Shadow, through its impact upon spectral curve shape and albedo, had a profound impact on classification using SAM and SVM.

Multichannel Pulse-Coupled-Neural-Network-Based Color Image Segmentation for Object Detection

This paper proposes a pulse-coupled neural network (PCNN) with multichannel (MPCNN) linking and feeding fields for color image segmentation. Different from the conventional PCNN, pulse-based radial basis function units are introduced into the model neurons of PCNN to determine the fast links among neurons with respect to their spectral feature vectors and spatial proximity. The computing of the color image segmentation can be implemented in parallel on a field-programmable-gate-array chip. Furthermore, the results of segmentations are applied to an object-detection scheme. Experimental results show that the performance of the proposed MPCNN is comparable to those of other popular image segmentation algorithms for the segmentation of noisy images while its parallel neural circuits improve the speed of processing drastically as compared with the sequential-code-based counterparts.

Multispectral texture characterization: application to computer aided diagnosis on prostatic tissue images

AbstractVarious approaches have been proposed in the literature for texture characterization of images. Some of them are based on statistical properties, others on fractal measures and some more on multi-resolution analysis. Basically, these approaches have been applied on mono-band images. However, most of them have been extended by including the additional information between spectral bands to deal with multi-band texture images. In this article, we investigate the problem of texture characterization for multi-band images. Therefore, we aim to add spectral information to classical texture analysis methods that only treat gray-level spatial variations. To achieve this goal, we propose a spatial and spectral gray level dependence method (SSGLDM) in order to extend the concept of gray level co-occurrence matrix (GLCM) by assuming the presence of texture joint information between spectral bands. Thus, we propose new multi-dimensional functions for estimating the second-order joint conditional probability density of spectral vectors. Theses functions can be represented in structure form which can help us to compute the occurrences while keeping the corresponding components of spectral vectors. In addition, new texture features measurements related to (SSGLDM) which define the multi-spectral image properties are proposed. Extensive experiments have been carried out on 624 textured multi-spectral images for use in prostate cancer diagnosis and quantitative results showed the efficiency of this method compared to the GLCM. The results indicate a significant improvement in terms of global accuracy rate. Thus, the proposed approach can provide clinically useful information for discriminating pathological tissue from healthy tissue.

Impact of Topography on Accuracy of Land Cover Spectral Change Vector Analysis Using AWiFS in Western Himalaya

The present paper discusses the impact of topography on accuracy for land cover classification and “from-to class change using improved spectral change vector analysis suggested by Chen et al. (2003). Two AWiFS sensor images of different dates are used. Double Window Flexible Pace Search (DFPS) is used to estimate threshold of change magnitude for change/no change classes. The topographic corrections show accuracy of 90% (Kappa coefficient 0.7811) for change/no change area as compared to 82% (Kappa coefficient 0.6512) in uncorrected satellite data. Direction cosines of change vector for determining change direction in n-dimensional spectral space is used for image classification with a minimum distance categorizing technique. The results of change detection are compared (i) Improved CVA with conventional two bands CVA and (ii) Improved CVA before and after topographic corrections. The improved CVA with topographic correction consideration using slope match show maximum accuracy of 90% (Kappa coefficient 0.83) as compared to conventional CVA which show maximum accuracy of 82% (Kappa coefficient 0.6624). The overall accuracy of ”from- to class using improved CVA increases from 86% (Kappa coefficient 0.7817) to 90% (Kappa coefficient 0.83) after topographic corrections. The improved CVA with proper topographic corrections is found to be effective for change detection analysis in the rugged Western Himalayan terrain.

Automatic recognition of frog calls using a multi-stage average spectrum

The automatic recognition of animal sounds is one of the powerful techniques for replacing the traditional ecological survey method that mainly depends on manpower, which is hence both costly and time consuming. This study developed an automatic frog call recognition system based on the combination of a pre-classification method of the syllable lengths and a multi-stage average spectrum (MSAS) method. In this system, the input frog syllables are first classified into one of the four groups determined by the pre-classification method according to syllable length. Then the proposed MSAS method is used to extract the standard feature template to analyze the time-varying features of each frog species and to recognize the input frog syllable by a template matching method. In all, 960 syllables recorded from 18 frog species are included in this study to evaluate the accuracy of the proposed frog call recognition system. The experimental results demonstrate that the proposed one-level (using the MSAS method only) and two-level (combining the syllable length pre-classification and MSAS methods) recognition methods can provide the best recognition accuracies of 91.9% and 94.3%, respectively, compared with other recognition methods based on dynamic time warping (DTW), spectral ensemble average voice prints (SEAV), k-nearest neighbor (kNN) and support vector machines (SVMs).

Ultrasound and Artificial Intelligence Applied to the Pollution Estimation in Insulations

This paper presents an electrical insulator pollution estimation technique based on the ultrasonic noise emitted by these insulators when connected to energized electrodes. The spectral subband centroid energy vectors (SSCEV) algorithm was employed in the signal processing. This algorithm can be understood as being a spectral compression, capable of selecting the most significant frequency bands. The processed audio, changed into SSCEV, constituted a database which was fed into an artificial neural network (ANN), capable of distinguishing with remarkable precision a SSCEV from a polluted insulator from a SSCEV from a less polluted insulator. Finally, in order to validate the technique in the field, measurement campaigns were performed at the Campina Grande 2 substation, belonging to the Sao Francisco Hydroelectric Company. During these campaigns, ultrasonic noise from several electrical equipments, exposed to different natural pollution degrees, was obtained, and the processing, based on SSCEV and the artificial neural network, was once again applied. As a result, the success rates of more than 80% were generally obtained by the ANN.

VOICE CONVERSION USING INTERPOLATED SPEECH UNIT START AND END−TIME CONVERSION RULE MATRICES AND SPECTRAL COMPENSATION ON ITS SPECTRAL PARAMETER VECTOR

A voice conversion rule and a rule selection parameter are stored. The voice conversion rule converts a spectral parameter vector of a source speaker to a spectral parameter vector of a target speaker. The rule selection parameter represents the spectral parameter vector of the source speaker. A first voice conversion rule of start time and a second voice conversion rule of end time in a speech unit of the source speaker are selected by the spectral parameter vector of the start time and the end time. An interpolation coefficient corresponding to the spectral parameter vector of each time in the speech unit is calculated by the first voice conversion rule and the second voice conversion rule. A third voice conversion rule corresponding to the spectral parameter vector of each time in the speech unit is calculated by interpolating the first voice conversion rule and the second voice conversion rule with the interpolation coefficient. The spectral parameter vector of each time is converted to a spectral parameter vector of the target speaker by the third voice conversion rule. A spectrum acquired from the spectral parameter vector of the target speaker is compensated by a spectral compensation filter or power ratio. A speech waveform is generated from the compensated spectrum.

Acoustic Modeling Using Deep Belief Networks

Gaussian mixture models are currently the dominant technique for modeling the emission distribution of hidden Markov models for speech recognition. We show that better phone recognition on the TIMIT dataset can be achieved by replacing Gaussian mixture models by deep neural networks that contain many layers of features and a very large number of parameters. These networks are first pre-trained as a multi-layer generative model of a window of spectral feature vectors without making use of any discriminative information. Once the generative pre-training has designed the features, we perform discriminative fine-tuning using backpropagation to adjust the features slightly to make them better at predicting a probability distribution over the states of monophone hidden Markov models.

Texture analysis of multi-spectral prostate tissue using Generalised Grey Level Difference Method

In this paper, we propose a new approach of multi-spectral texture classification based on both spatial and spectral information. We do this by extending the concept of spatial Grey Level Difference Method (GLDM) and assuming texture joint information between spectral bands. In this manner, we can characterise multi-spectral texture by statistics of absolute difference distributions of pairs of spectral vectors and define new texture features by computing various statistics from such distributions in given relative positions. Extensive experiments have been carried out on many multi-spectral images for use in prostate cancer diagnosis and quantitative results showed the efficiency of this method compared with the GLDM. The results indicate a significant improvement in terms of global accuracy rate.

Improving the classification of brain tumors in mice with perturbation enhanced (PE)-MRSI

Classifiers based on statistical pattern recognition analysis of MRSI data are becoming important tools for the non-invasive diagnosis of human brain tumors. Here we investigate the potential interest of perturbation-enhanced MRSI (PE-MRSI), in this case acute hyperglycemia, for improving the discrimination between mouse brain MRS patterns of glioblastoma multiforme (GBM), oligodendroglioma (ODG), and non-tumor brain parenchyma (NT). Six GBM-bearing mice and three ODG-bearing mice were scanned at 7 Tesla by PRESS-MRSI with 12 and 136 ms echo-time, during euglycemia (Eug) and also during induced acute hyperglycemia (Hyp), generating altogether four datasets per animal (echo time + glycemic condition): 12Eug, 136Eug, 12Hyp, and 136Hyp. For classifier development all spectral vectors (spv) selected from the MRSI matrix were unit length normalized (UL2) and used either as a training set (76 GBM spv, four mice; 70 ODG spv, two mice; 54 NT spv) or as an independent testing set (61 GBM spv, two mice; 31 ODG, one mouse; 23 NT spv). All Fisher's LDA classifiers obtained were evaluated as far as their descriptive performance-correctly classified cases of the training set (bootstrapping)-and predictive accuracy-balanced error rate of independent testing set classification. MRSI-based classifiers at 12Hyp were consistently more efficient in separating GBM, ODG, and NT regions, with overall accuracies always >80% and up to 95-96%; remaining classifiers were within the 48-85% range. This was also confirmed by user-independent selection of training and testing sets, using leave-one-out (LOO). This highlights the potential interest of perturbation-enhanced MRSI protocols for improving the non-invasive characterization of preclinical brain tumors.

Stability indicating analysis of bisacodyl by partial least squares regression, spectral residual augmented classical least squares and support vector regression chemometric models: A comparative study

Partial least squares regression (PLSR), spectral residual augmented classical least squares (SRACLS) and support vector regression (SVR) are three different chemometric models. These models are subjected to a comparative study that highlights their inherent characteristics via applying them to analysis of bisacodyl in the presence of its reported degradation products monoacetyl bisacodyl (I) and desacetyl bisacodyl (II), in raw material. For proper analysis, a 3 factor 3 level experimental design was established resulting in a training set of 9 mixtures containing different ratios of the interfering species. A linear test set consisting of 6 mixtures was used to validate the prediction ability of the suggested models. To test the generalisation ability of the models, some extra mixtures were prepared that are outside the concentration space of the training set. To test the ability of models to handle nonlinearity in spectral response, another set of nonlinear samples was prepared. The paper highlights model transfer to other labs under other conditions as well. This paper aims to manifest the advantages of SRACLS and SVR over PLSR model, where SRACLS can tackle future changes without the need for tedious recalibration, while SVR is a more robust and general model, with high ability to model nonlinearity in spectral response, though like PLSR is needing recalibration. The results presented indicate the ability of the three models to analyse bisacodyl in the presence of its degradation products in raw material with high accuracy and precision; where SVR gives the best results at all tested conditions compared to other models.

Histogram thresholding for unsupervised change detection of remote sensing images

The change-detection problem can be viewed as an unsupervised classification problem with two classes corresponding to changed and unchanged areas. Image differencing is a widely used approach to change detection. It is based on the idea of generating a difference image that represents the modulus of the spectral change vectors associated with each pixel in the study area. To separate out the changed and unchanged classes in the difference image automatically, any unsupervised technique can be used. Thresholding is one of the cheapest techniques among them. However, in thresholding approaches, selection of the best threshold value is not a trivial task. In this work, several non-fuzzy and fuzzy histogram thresholding techniques are investigated and compared for the change-detection problem. Experimental results, carried out on different multitemporal remote sensing images (acquired before and after an event), are used to assess the effectiveness of each of the thresholding techniques. Among all the thresholding techniques investigated here, Liu's fuzzy entropy followed by Kapur's entropy are found to be the most robust techniques.

Change Detection from Landsat TM Images Using Change Vector Analysis

Change detection is a process of extracting, analyzing, and defining change information from remote sensing imageries. At present, remote sensing change detection methods are mainly classified into two categories, one based on the spectral characteristics of the type of approach, and the other is spectral change vector analysis. In this paper, a simplified threshold variable step-size search which can be used on determining changes in the threshold vector is adopted, as well as the supervised classification technique in the direction cosine space with the type of focal point in the initial assay vector remote sensing images. Results are discussed in the last part of this paper, which show that CVA can extract change information effectively in our study area of Wuhan city.

Bayesian Hyperspectral Image Segmentation With Discriminative Class Learning

This paper introduces a new supervised technique to segment hyperspectral images: the Bayesian segmentation based on discriminative classification and on multilevel logistic (MLL) spatial prior. The approach is Bayesian and exploits both spectral and spatial information. Given a spectral vector, the posterior class probability distribution is modeled using multinomial logistic regression (MLR) which, being a discriminative model, allows to learn directly the boundaries between the decision regions and, thus, to successfully deal with high-dimensionality data. To control the machine complexity and, thus, its generalization capacity, the prior on the multinomial logistic vector is assumed to follow a componentwise independent Laplacian density. The vector of weights is computed via the fast sparse multinomial logistic regression (FSMLR), a variation of the sparse multinomial logistic regression (SMLR), conceived to deal with large data sets beyond the reach of the SMLR. To avoid the high computational complexity involved in estimating the Laplacian regularization parameter, we have also considered the Jeffreys prior, as it does not depend on any hyperparameter. The prior probability distribution on the class-label image is an MLL Markov–Gibbs distribution, which promotes segmentation results with equal neighboring class labels. The <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$\alpha$</tex></formula> -expansion optimization algorithm, a powerful graph-cut-based integer optimization tool, is used to compute the maximum a posteriori segmentation. The effectiveness of the proposed methodology is illustrated by comparing its performance with the state-of-the-art methods on synthetic and real hyperspectral image data sets. The reported results give clear evidence of the relevance of using both spatial and spectral information in hyperspectral image segmentation.

Mapping Malaysian Urban Environment from Airborne Hyperspectral Sensor System in the VIS-NIR (0.4-1.1 µm) Spectrum

Airborne hyperspectral remote sensing is a relatively new technology in Malaysia that needs to be tested for its feasibility. Various applications can benefit from the enormous potential offered such as in urban mapping in which rapid development in Malaysia can be accurately monitored. However, the use of hyperspectral data will also depend critically on the selection of suitable classifiers in order to extract the information. Hence, in this study, image classification was performed using various classifiers such as Parallelepiped, Minimum Distance, Mahalanobis Distance, Maximum Likelihood (ML), Spectral Information Divergence (SID), Spectral Angle Mapper (SAM), Binary Encoding (BE), Neural Network (NN) and Support Vector Machine (SVM). The accuracy of the classifiers was measured based on comparisons with ground truth data. SVM classifier shows the highest overall accuracy (87.98%) followed by ML with 83.17% and BE achieved the lowest accuracy with 39.28%. The findings indicate the feasibility of hyperspectral remote sensing for mapping urban environment in Malaysia with SVM as the most effective classifier for that purpose.

Spectral change directions of multispectral subpixels in image fusion

In an image fusion process, the spatial detail in a panchromatic pixel is injected into the corresponding n-band multispectral (MS) subpixel to yield a synthetic pixel. The synthetic pixel can be regarded as the sum of three terms: the MS subpixel, a shift term, and a product of the spatial detail and an n-dimensional (n-D) spectral change vector. In this paper, the spectral change vector directions in some current image fusion methods are characterized and classified. As the spectral development of subpixels with the improvements in spatial resolution in imaging is analogous to the spectral change of MS subpixels with the improvements in spatial resolution in image fusion, the former is used as a reference to examine the spectral change vector directions in the current image fusion methods. Moreover, image haze and unmixing of mixed MS subpixels are highlighted as two aspects needing attention for further improvement in fusion quality.

Spectral requirements on airborne hyperspectral remote sensing data for wheat disease detection

Remote sensing approaches are of increasing importance for agricultural applications, particularly for the support of selective agricultural measures that increase the productivity of crop stands. In contrast to multi-spectral image data, hyperspectral data has been shown to be highly suitable for the detection of crop growth anomalies, since they allow a detailed examination of stress-dependent changes in certain spectral ranges. However, the entire spectrum covered by hyperspectral data is probably not needed for discrimination between healthy and stressed plants. To define an optimal sensor-based system or a data product designed for crop stress detection, it is necessary to know which spectral wavelengths are significantly affected by stress factors and which spectral resolution is needed. In this study, a single airborne hyperspectral HyMap dataset was analyzed for its potential to detect plant stress symptoms in wheat stands induced by a pathogen infection. The Bhattacharyya distance (BD) with a forward feature search strategy was used to select relevant bands for the differentiation between healthy and fungal infected stands. Two classification algorithms, i.e. spectral angle mapper (SAM) and support vector machines (SVM) were used to classify the data covering an experimental field. Thus, the original dataset as well as datasets reduced to several band combinations as selected by the feature selection approach were classified. To analyze the influence of the spectral resolution on the detection accuracy, the original dataset was additionally stepwise spectrally resampled and a feature selection was carried out on each step. It is demonstrated that just a few phenomenon-specific spectral features are sufficient to detect wheat stands infected with powdery mildew. With original spectral resolution of HyMap, the highest classification accuracy could be obtained by using only 13 spectral bands with a Kappa coefficient of 0.59 in comparison to Kappa 0.57 using all spectral bands of the HyMap sensor. The results demonstrate that even a few hyperspectral bands as well as bands with lower spectral resolution still allow an adequate detection of fungal infections in wheat. By focusing on a few relevant bands, the detection accuracy could be enhanced and thus more reliable information could be extracted which may be helpful in agricultural practice.

A Multifeature Tensor for Remote-Sensing Target Recognition

In remote-sensing image target recognition, the target or background object is usually transformed to a feature vector, such as a spectral feature vector. However, this kind of vector represents only one pixel of a remote-sensing image that considers the spectral information but ignores the spatial relationship of neighboring pixels (i.e., the local texture and structure). In this letter, we propose a new way to represent an image object as a multifeature tensor that encodes both the spectral and textural information (Gabor function) and then apply the support tensor machine for target recognition. A range of experiments demonstrates that the effectiveness of the proposed method can deliver a high and correct recognition rate with a small number of training samples.