Unmixing hyperspectral data: independent and dependent component analysis

The development of high spatial resolution airborne and spaceborne sensors has improved the capability of ground-based data collection in the fields of agriculture, geography, geology, mineral identification, detection [2, 3], and classification [4–8]. The signal read by the sensor from a given spatial element of resolution and at a given spectral band is a mixing of components originated by the constituent substances, termed endmembers, located at that element of resolution. This chapter addresses hyperspectral unmixing, which is the decomposition of the pixel spectra into a collection of constituent spectra, or spectral signatures, and their corresponding fractional abundances indicating the proportion of each endmember present in the pixel [9, 10]. Depending on the mixing scales at each pixel, the observed mixture is either linear or nonlinear [11, 12]. The linear mixing model holds when the mixing scale is macroscopic [13]. The nonlinear model holds when the mixing scale is microscopic (i.e., intimate mixtures) [14, 15]. The linear model assumes negligible interaction among distinct endmembers [16, 17]. The nonlinear model assumes that incident solar radiation is scattered by the scene through multiple bounces involving several endmembers [18]. Under the linear mixing model and assuming that the number of endmembers and their spectral signatures are known, hyperspectral unmixing is a linear problem, which can be addressed, for example, under the maximum likelihood setup [19], the constrained least-squares approach [20], the spectral signature matching [21], the spectral angle mapper [22], and the subspace projection methods [20, 23, 24]. Orthogonal subspace projection [23] reduces the data dimensionality, suppresses undesired spectral signatures, and detects the presence of a spectral signature of interest. The basic concept is to project each pixel onto a subspace that is orthogonal to the undesired signatures. As shown in Settle [19], the orthogonal subspace projection technique is equivalent to the maximum likelihood estimator. This projection technique was extended by three unconstrained least-squares approaches [24] (signature space orthogonal projection, oblique subspace projection, target signature space orthogonal projection). Other works using maximum a posteriori probability (MAP) framework [25] and projection pursuit [26, 27] have also been applied to hyperspectral data. In most cases the number of endmembers and their signatures are not known. Independent component analysis (ICA) is an unsupervised source separation process that has been applied with success to blind source separation, to feature extraction, and to unsupervised recognition [28, 29]. ICA consists in finding a linear decomposition of observed data yielding statistically independent components. Given that hyperspectral data are, in given circumstances, linear mixtures, ICA comes to mind as a possible tool to unmix this class of data. In fact, the application of ICA to hyperspectral data has been proposed in reference 30, where endmember signatures are treated as sources and the mixing matrix is composed by the abundance fractions, and in references 9, 25, and 31–38, where sources are the abundance fractions of each endmember. In the first approach, we face two problems: (1) The number of samples are limited to the number of channels and (2) the process of pixel selection, playing the role of mixed sources, is not straightforward. In the second approach, ICA is based on the assumption of mutually independent sources, which is not the case of hyperspectral data, since the sum of the abundance fractions is constant, implying dependence among abundances. This dependence compromises ICA applicability to hyperspectral images. In addition, hyperspectral data are immersed in noise, which degrades the ICA performance. IFA [39] was introduced as a method for recovering independent hidden sources from their observed noisy mixtures. IFA implements two steps. First, source densities and noise covariance are estimated from the observed data by maximum likelihood. Second, sources are reconstructed by an optimal nonlinear estimator. Although IFA is a well-suited technique to unmix independent sources under noisy observations, the dependence among abundance fractions in hyperspectral imagery compromises, as in the ICA case, the IFA performance. Considering the linear mixing model, hyperspectral observations are in a simplex whose vertices correspond to the endmembers. Several approaches [40–43] have exploited this geometric feature of hyperspectral mixtures [42]. Minimum volume transform (MVT) algorithm [43] determines the simplex of minimum volume containing the data. The MVT-type approaches are complex from the computational point of view. Usually, these algorithms first find the convex hull defined by the observed data and then fit a minimum volume simplex to it. Aiming at a lower computational complexity, some algorithms such as the vertex component analysis (VCA) [44], the pixel purity index (PPI) [42], and the N-FINDR [45] still find the minimum volume simplex containing the data cloud, but they assume the presence in the data of at least one pure pixel of each endmember. This is a strong requisite that may not hold in some data sets. In any case, these algorithms find the set of most pure pixels in the data. Hyperspectral sensors collects spatial images over many narrow contiguous bands, yielding large amounts of data. For this reason, very often, the processing of hyperspectral data, included unmixing, is preceded by a dimensionality reduction step to reduce computational complexity and to improve the signal-to-noise ratio (SNR). Principal component analysis (PCA) [46], maximum noise fraction (MNF) [47], and singular value decomposition (SVD) [48] are three well-known projection techniques widely used in remote sensing in general and in unmixing in particular. The newly introduced method [49] exploits the structure of hyperspectral mixtures, namely the fact that spectral vectors are nonnegative. The computational complexity associated with these techniques is an obstacle to real-time implementations. To overcome this problem, band selection [50] and non-statistical [51] algorithms have been introduced. This chapter addresses hyperspectral data source dependence and its impact on ICA and IFA performances. The study consider simulated and real data and is based on mutual information minimization. Hyperspectral observations are described by a generative model. This model takes into account the degradation mechanisms normally found in hyperspectral applications—namely, signature variability [52–54], abundance constraints, topography modulation, and system noise. The computation of mutual information is based on fitting mixtures of Gaussians (MOG) to data. The MOG parameters (number of components, means, covariances, and weights) are inferred using the minimum description length (MDL) based algorithm [55]. We study the behavior of the mutual information as a function of the unmixing matrix. The conclusion is that the unmixing matrix minimizing the mutual information might be very far from the true one. Nevertheless, some abundance fractions might be well separated, mainly in the presence of strong signature variability, a large number of endmembers, and high SNR. We end this chapter by sketching a new methodology to blindly unmix hyperspectral data, where abundance fractions are modeled as a mixture of Dirichlet sources. This model enforces positivity and constant sum sources (full additivity) constraints. The mixing matrix is inferred by an expectation-maximization (EM)-type algorithm. This approach is in the vein of references 39 and 56, replacing independent sources represented by MOG with mixture of Dirichlet sources. Compared with the geometric-based approaches, the advantage of this model is that there is no need to have pure pixels in the observations. The chapter is organized as follows. Section 6.2 presents a spectral radiance model and formulates the spectral unmixing as a linear problem accounting for abundance constraints, signature variability, topography modulation, and system noise. Section 6.3 presents a brief resume of ICA and IFA algorithms. Section 6.4 illustrates the performance of IFA and of some well-known ICA algorithms with experimental data. Section 6.5 studies the ICA and IFA limitations in unmixing hyperspectral data. Section 6.6 presents results of ICA based on real data. Section 6.7 describes the new blind unmixing scheme and some illustrative examples. Section 6.8 concludes with some remarks.

Vertex component analysis: a geometric-based approach to unmix hyperspectral data

Hyperspectral remote sensing exploits the electromagnetic scattering patterns of the different materials at specific wavelengths [2, 3]. Hyperspectral sensors have been developed to sample the scattered portion of the electromagnetic spectrum extending from the visible region through the near-infrared and mid-infrared, in hundreds of narrow contiguous bands [4, 5]. The number and variety of potential civilian and military applications of hyperspectral remote sensing is enormous [6, 7]. Very often, the resolution cell corresponding to a single pixel in an image contains several substances (endmembers) [4]. In this situation, the scattered energy is a mixing of the endmember spectra. A challenging task underlying many hyperspectral imagery applications is then decomposing a mixed pixel into a collection of reflectance spectra, called endmember signatures, and the corresponding abundance fractions [8–10]. Depending on the mixing scales at each pixel, the observed mixture is either linear or nonlinear [11, 12]. Linear mixing model holds approximately when the mixing scale is macroscopic [13] and there is negligible interaction among distinct endmembers [3, 14]. If, however, the mixing scale is microscopic (or intimate mixtures) [15, 16] and the incident solar radiation is scattered by the scene through multiple bounces involving several endmembers [17], the linear model is no longer accurate. Linear spectral unmixing has been intensively researched in the last years [9, 10, 12, 18–21]. It considers that a mixed pixel is a linear combination of endmember signatures weighted by the correspondent abundance fractions. Under this model, and assuming that the number of substances and their reflectance spectra are known, hyperspectral unmixing is a linear problem for which many solutions have been proposed (e.g., maximum likelihood estimation [8], spectral signature matching [22], spectral angle mapper [23], subspace projection methods [24,25], and constrained least squares [26]). In most cases, the number of substances and their reflectances are not known and, then, hyperspectral unmixing falls into the class of blind source separation problems [27]. Independent component analysis (ICA) has recently been proposed as a tool to blindly unmix hyperspectral data [28–31]. ICA is based on the assumption of mutually independent sources (abundance fractions), which is not the case of hyperspectral data, since the sum of abundance fractions is constant, implying statistical dependence among them. This dependence compromises ICA applicability to hyperspectral images as shown in Refs. [21, 32]. In fact, ICA finds the endmember signatures by multiplying the spectral vectors with an unmixing matrix, which minimizes the mutual information among sources. If sources are independent, ICA provides the correct unmixing, since the minimum of the mutual information is obtained only when sources are independent. This is no longer true for dependent abundance fractions. Nevertheless, some endmembers may be approximately unmixed. These aspects are addressed in Ref. [33]. Under the linear mixing model, the observations from a scene are in a simplex whose vertices correspond to the endmembers. Several approaches [34–36] have exploited this geometric feature of hyperspectral mixtures [35]. Minimum volume transform (MVT) algorithm [36] determines the simplex of minimum volume containing the data. The method presented in Ref. [37] is also of MVT type but, by introducing the notion of bundles, it takes into account the endmember variability usually present in hyperspectral mixtures. The MVT type approaches are complex from the computational point of view. Usually, these algorithms find in the first place the convex hull defined by the observed data and then fit a minimum volume simplex to it. For example, the gift wrapping algorithm [38] computes the convex hull of n data points in a d-dimensional space with a computational complexity of O(nbd=2cþ1), where bxc is the highest integer lower or equal than x and n is the number of samples. The complexity of the method presented in Ref. [37] is even higher, since the temperature of the simulated annealing algorithm used shall follow a log( ) law [39] to assure convergence (in probability) to the desired solution. Aiming at a lower computational complexity, some algorithms such as the pixel purity index (PPI) [35] and the N-FINDR [40] still find the minimum volume simplex containing the data cloud, but they assume the presence of at least one pure pixel of each endmember in the data. This is a strong requisite that may not hold in some data sets. In any case, these algorithms find the set of most pure pixels in the data. PPI algorithm uses the minimum noise fraction (MNF) [41] as a preprocessing step to reduce dimensionality and to improve the signal-to-noise ratio (SNR). The algorithm then projects every spectral vector onto skewers (large number of random vectors) [35, 42,43]. The points corresponding to extremes, for each skewer direction, are stored. A cumulative account records the number of times each pixel (i.e., a given spectral vector) is found to be an extreme. The pixels with the highest scores are the purest ones. N-FINDR algorithm [40] is based on the fact that in p spectral dimensions, the p-volume defined by a simplex formed by the purest pixels is larger than any other volume defined by any other combination of pixels. This algorithm finds the set of pixels defining the largest volume by inflating a simplex inside the data. ORA SIS [44, 45] is a hyperspectral framework developed by the U.S. Naval Research Laboratory consisting of several algorithms organized in six modules: exemplar selector, adaptative learner, demixer, knowledge base or spectral library, and spatial postrocessor. The first step consists in flat-fielding the spectra. Next, the exemplar selection module is used to select spectral vectors that best represent the smaller convex cone containing the data. The other pixels are rejected when the spectral angle distance (SAD) is less than a given thresh old. The procedure finds the basis for a subspace of a lower dimension using a modified Gram–Schmidt orthogonalizati on. The selected vectors are then projected onto this subspace and a simplex is found by an MV T pro cess. ORA SIS is oriented to real-time target detection from uncrewed air vehicles using hyperspectral data [46]. In this chapter we develop a new algorithm to unmix linear mixtures of endmember spectra. First, the algorithm determines the number of endmembers and the signal subspace using a newly developed concept [47, 48]. Second, the algorithm extracts the most pure pixels present in the data. Unlike other methods, this algorithm is completely automatic and unsupervised. To estimate the number of endmembers and the signal subspace in hyperspectral linear mixtures, the proposed scheme begins by estimating sign al and noise correlation matrices. The latter is based on multiple regression theory. The signal subspace is then identified by selectin g the set of signal eigenvalue s that best represents the data, in the least-square sense [48,49 ], we note, however, that VCA works with projected and with unprojected data. The extraction of the end members exploits two facts: (1) the endmembers are the vertices of a simplex and (2) the affine transformation of a simplex is also a simplex. As PPI and N-FIND R algorithms, VCA also assumes the presence of pure pixels in the data. The algorithm iteratively projects data on to a direction orthogonal to the subspace spanned by the endmembers already determined. The new end member signature corresponds to the extreme of the projection. The algorithm iterates until all end members are exhausted. VCA performs much better than PPI and better than or comparable to N-FI NDR; yet it has a computational complexity between on e and two orders of magnitude lower than N-FINDR. The chapter is structure d as follows. Section 19.2 describes the fundamentals of the proposed method. Section 19.3 and Section 19.4 evaluate the proposed algorithm using simulated and real data, respectively. Section 19.5 presents some concluding remarks.

Utilização de métodos ópticos de correlação de imagem para a determinação da tenacidade à fractura de adesivos estruturais

Mestrado em Engenharia Mecânica

The value of adenosine deaminase (ADA) determination in the diagnosis of tuberculous ascites

In order to evaluate the role of the determination of adenosine deaminase activity (ADA) in ascitic fluid for the diagnosis of tuberculosis, 44 patients were studied. Based on biochemical, cytological, histopathological and microbiological tests, the patients were divided into 5 groups: G1 - tuberculous ascites (n = 8); G2 - malignant ascites (n = 13); G3 - spontaneous bacterial peritonitis (n = 6); G4 - pancreatic ascites (n = 2); G5 - miscelaneous ascites (n = 15). ADA concentration were significantly higher in G1 (133.50 ± 24.74 U/l) compared to the other groups (G2 = 41.85 ± 52.07 U/l; G3 = 10.63 ± 5.87 U/l; G4 = 18.00 ± 7.07 U/l; G5 = 11.23 ± 7.66 U/l). At a cut-off value of >31 U/l, the sensitivity, specificity and positive and negative preditive values were 100%, 92%, 72% and 100%, respectively. ADA concentrations as high as in tuberculous ascites were only found in two malignant ascites caused by lymphoma. We conclude that ADA determination in ascitic fluid is a useful and reliable screening test for diagnosing tuberculous ascites. Values of ADA higher than 31 U/l indicate more invasive methods to confirm the diagnosis of tuberculosis.

Determinants of frailty: the added value of assessing medication

This study aims to analyze which determinants predict frailty in general and each frailty domain (physical, psychological, and social), considering the integral conceptual model of frailty, and particularly to examine the contribution of medication in this prediction. A cross-sectional study was designed using a non-probabilistic sample of 252 community-dwelling elderly from three Portuguese cities. Frailty and determinants of frailty were assessed with the Tilburg Frailty Indicator. The amount and type of different daily-consumed medication were also examined. Hierarchical regression analysis were conducted. The mean age of the participants was 79.2 years (±7.3), and most of them were women (75.8%), widowed (55.6%) and with a low educational level (0–4 years: 63.9%). In this study, determinants explained 46% of the variance of total frailty, and 39.8, 25.3, and 27.7% of physical, psychological, and social frailty respectively. Age, gender, income, death of a loved one in the past year, lifestyle, satisfaction with living environment and self-reported comorbidity predicted total frailty, while each frailty domain was associated with a different set of determinants. The number of daily-consumed drugs was independently associated with physical frailty, and the consumption of medication for the cardiovascular system and for the blood and blood-forming organs explained part of the variance of total and physical frailty. The adverse effects of polymedication and its direct link with the level of comorbidities could explain the independent contribution of the amount of prescribed drugs to frailty prediction. On the other hand, findings in regard to medication type provide further evidence of the association of frailty with cardiovascular risk. In the present study, a significant part of frailty was predicted, and the different contributions of each determinant to frailty domains highlight the relevance of the integral model of frailty. The added value of a simple assessment of medication was considerable, and it should be taken into account for effective identification of frailty.

Strategic Bidding in Electricity Markets: An agent-based simulator with game theory for scenario analysis

Electricity markets are complex environments, involving a large number of different entities, with specific characteristics and objectives, making their decisions and interacting in a dynamic scene. Game-theory has been widely used to support decisions in competitive environments; therefore its application in electricity markets can prove to be a high potential tool. This paper proposes a new scenario analysis algorithm, which includes the application of game-theory, to evaluate and preview different scenarios and provide players with the ability to strategically react in order to exhibit the behavior that better fits their objectives. This model includes forecasts of competitor players’ actions, to build models of their behavior, in order to define the most probable expected scenarios. Once the scenarios are defined, game theory is applied to support the choice of the action to be performed. Our use of game theory is intended for supporting one specific agent and not for achieving the equilibrium in the market. MASCEM (Multi-Agent System for Competitive Electricity Markets) is a multi-agent electricity market simulator that models market players and simulates their operation in the market. The scenario analysis algorithm has been tested within MASCEM and our experimental findings with a case study based on real data from the Iberian Electricity Market are presented and discussed.

Every normal logic program has a 2-valued semantics: theory, extensions, applications, implementations

Trabalho apresentado no âmbito do Doutoramento em Informática, como requisito parcial para obtenção do grau de Doutor em Informática

Health care marketing: the theory of planned behavior applied to patients' choise between private and public providers in the portuguese health care system.

Dissertation submitted in partial fulfillment of the requirements for degree of Master in Statistics and Information Management.

Comparação dos ensaios DCB e TDCB para determinação da tenacidade em tração de adesivos estruturais

O constante desenvolvimento observado nas ligações adesivas, em conjunto com as melhorias verificadas nas características dos adesivos, estão a ser traduzidos, de uma certa forma, num aumento das aplicações das ligações adesivas, assim como na variedade de aplicações. No âmbito da previsão de resistência de juntas adesivas, dois métodos de grande relevância são a Mecânica da Fratura e os Modelos de Dano Coesivo. Os Modelos de Dano Coesivo permitem a simulação da iniciação e propagação do dano, recorrendo ao Método Dos Elementos Finitos. No que concerne à Mecânica da Fratura, a previsão de resistência é geralmente feita através de uma análise energética. Independentemente da forma como é obtida, a taxa crítica de libertação de energia de deformação à tração (GIc) dos adesivos é um dos parâmetros mais importantes para a previsão da resistência das juntas. Dois dos ensaios mais utilizados são o Double Cantilever Beam (DCB) e o Tapered Double Cantilever Beam (TDCB). Este trabalho pretende determinar e comparar o valor de GIc em juntas adesivas pelos ensaios DCB e TDCB. São utilizados três tipos de adesivos com diferentes graus de ductilidade. No ensaio DCB os métodos utilizados para a determinação de GIc são: Compliance-Based Beam Method (CBBM), Corrected Beam Theory (CBT) e Compliance Calibration Method (CCM). Os métodos utilizados no ensaio TDCB são: Simple Beam Theory (SBT), Corrected Beam Theory (CBT) e Compliance Calibration Method (CCM). Os resultados obtidos apresentam concordância entre os vários métodos de cada ensaio. A discrepância de resultados é superior quando comparados os dois tipos de ensaios.

Depositions of rulers in the later middle ages: on the theory of th "useless ruler" and its practical utilization

pp. 157-168

Bipedal Locomotion: A Fractional CPG for Generating Patterns

Proceedings of the 10th Conference on Dynamical Systems Theory and Applications

“The past is a foreign country”? Acculturation theory and the anthropology of globalization

Etnográfica, 15 (2): 313-336

Teorias e modelos organizacionais

Este trabalho propõe-se a investigar as teorias e modelos organizacionais e a respetiva aplicabilidade nas organizações portuguesas. Após a revisão da literatura sobre modelos organizacionais, foi efetuada uma investigação quantitativa através de um questionário online com a finalidade de avaliar quais os modelos organizacionais predominantemente utilizados e quais as características organizacionais que levam à utilização de determinado modelo. Através de métodos estatísticos analisaram-se os resultados do inquérito com o objetivo de verificar a existência de possíveis relações entre diversas características das organizações e o modelo organizacional usado. Foi possível concluir que o modelo organizacional Burocrático é o modelo predominantemente utilizado pelos respondentes e que as organizações que adotam o modelo burocrático parecem conseguir implementar processos sistemáticos de inovação compatibilizando as regras e procedimentos com a capacidade para aprender e se adaptar. O setor de atividade e a dimensão das organizações são as variáveis que mais influenciam a adoção do modelo organizacional. A investigação contribui para o conhecimento teórico e pratico sobre modelos organizacionais e sobre a sua aplicação em diferentes tipos de organizações portuguesas e para a compreensão e capacitação dos engenheiros do tema da cultura organizacional, de modo a poderem trabalhar de forma efetiva em grupos multidisciplinares que criem valor para as respetivas organizações, inovando e aplicando a engenharia e tecnologia para lidar com as questões e desafios atuais referidos pelo relatório da UNESCO (1).

Prognostic Value of a New Cardiopulmonary Exercise Testing Parameter in Chronic Heart Failure: Oxygen Uptake Efficiency at Peak Exercise - Comparison with Oxygen Uptake Efficiency Slope

INTRODUCTION: A growing body of evidence shows the prognostic value of oxygen uptake efficiency slope (OUES), a cardiopulmonary exercise test (CPET) parameter derived from the logarithmic relationship between O(2) consumption (VO(2)) and minute ventilation (VE) in patients with chronic heart failure (CHF). OBJECTIVE: To evaluate the prognostic value of a new CPET parameter - peak oxygen uptake efficiency (POUE) - and to compare it with OUES in patients with CHF. METHODS: We prospectively studied 206 consecutive patients with stable CHF due to dilated cardiomyopathy - 153 male, aged 53.3±13.0 years, 35.4% of ischemic etiology, left ventricular ejection fraction 27.7±8.0%, 81.1% in sinus rhythm, 97.1% receiving ACE-Is or ARBs, 78.2% beta-blockers and 60.2% spironolactone - who performed a first maximal symptom-limited treadmill CPET, using the modified Bruce protocol. In 33% of patients an cardioverter-defibrillator (ICD) or cardiac resynchronization therapy device (CRT-D) was implanted during follow-up. Peak VO(2), percentage of predicted peak VO(2), VE/VCO(2) slope, OUES and POUE were analyzed. OUES was calculated using the formula VO(2) (l/min) = OUES (log(10)VE) + b. POUE was calculated as pVO(2) (l/min) / log(10)peakVE (l/min). Correlation coefficients between the studied parameters were obtained. The prognosis of each variable adjusted for age was evaluated through Cox proportional hazard models and R2 percent (R2%) and V index (V6) were used as measures of the predictive accuracy of events of each of these variables. Receiver operating characteristic (ROC) curves from logistic regression models were used to determine the cut-offs for OUES and POUE. RESULTS: pVO(2): 20.5±5.9; percentage of predicted peak VO(2): 68.6±18.2; VE/VCO(2) slope: 30.6±8.3; OUES: 1.85±0.61; POUE: 0.88±0.27. During a mean follow-up of 33.1±14.8 months, 45 (21.8%) patients died, 10 (4.9%) underwent urgent heart transplantation and in three patients (1.5%) a left ventricular assist device was implanted. All variables proved to be independent predictors of this combined event; however, VE/VCO2 slope was most strongly associated with events (HR 11.14). In this population, POUE was associated with a higher risk of events than OUES (HR 9.61 vs. 7.01), and was also a better predictor of events (R2: 28.91 vs. 22.37). CONCLUSION: POUE was more strongly associated with death, urgent heart transplantation and implantation of a left ventricular assist device and proved to be a better predictor of events than OUES. These results suggest that this new parameter can increase the prognostic value of CPET in patients with CHF.

Coalition of distributed generation units to Virtual Power Players - a game theory approach

Smart Grids (SGs) have emerged as the new paradigm for power system operation and management, being designed to include large amounts of distributed energy resources. This new paradigm requires new Energy Resource Management (ERM) methodologies considering different operation strategies and the existence of new management players such as several types of aggregators. This paper proposes a methodology to facilitate the coalition between distributed generation units originating Virtual Power Players (VPP) considering a game theory approach. The proposed approach consists in the analysis of the classifications that were attributed by each VPP to the distributed generation units, as well as in the analysis of the previous established contracts by each player. The proposed classification model is based in fourteen parameters including technical, economical and behavioural ones. Depending of the VPP strategies, size and goals, each parameter has different importance. VPP can also manage other type of energy resources, like storage units, electric vehicles, demand response programs or even parts of the MV and LV distribution network. A case study with twelve VPPs with different characteristics and one hundred and fifty real distributed generation units is included in the paper.