Konstantinos Koutroumbas

National Observatory of Athens, Athínai, Attica, Greece

Are you Konstantinos Koutroumbas?

Claim your profile

Publications (39)26.28 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: In a plethora of applications dealing with inverse problems, e.g. in image processing, social networks, compressive sensing, biological data processing etc., the signal of interest is known to be structured in several ways at the same time. This premise has recently guided the research to the innovative and meaningful idea of imposing multiple constraints on the parameters involved in the problem under study. For instance, when dealing with problems whose parameters form sparse and low-rank matrices, the adoption of suitably combined constraints imposing sparsity and low-rankness, is expected to yield substantially enhanced estimation results. In this paper, we address the spectral unmixing problem in hyperspectral images. Specifically, two novel unmixing algorithms are introduced, in an attempt to exploit both spatial correlation and sparse representation of pixels lying in homogeneous regions of hyperspectral images. To this end, a novel convex mixed penalty term is first defined consisting of the sum of the weighted $\ell_1$ and the weighted nuclear norm of the abundance matrix corresponding to a small area of the image determined by a sliding square window. This penalty term is then used to regularize a conventional quadratic cost function and impose simultaneously sparsity and row-rankness on the abundance matrix. The resulting regularized cost function is minimized by a) an incremental proximal sparse and low-rank unmixing algorithm and b) an algorithm based on the alternating minimization method of multipliers (ADMM). The effectiveness of the proposed algorithms is illustrated in experiments conducted both on simulated and real data.
  • Source
    Spyridoula D. Xenaki, Konstantinos D. Koutroumbas, Athanasios A. Rontogiannis
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper a novel possibilistic c-means clustering algorithm, called Adaptive Possibilistic c-means, is presented. Its main feature is that {\it all} its parameters, after their initialization, are properly adapted during its execution. This increases the flexibility of the algorithm in following the variations in the formation of the clusters that occur from iteration to iteration. The fully adaptive nature of the proposed algorithm enables the removal of the clusters that gradually become obsolete. This makes the algorithm capable to determine, in principle, the true number of the clusters underlying the data set (a long-standing issue in the clustering literature), provided that it starts with a reasonable overestimate of the actual number of clusters. However, the algorithm necessitates an appropriate, but fully determined in most cases, scaling of the data set before its execution. Extensive simulation results on both synthetic and real data highlight the effectiveness of the proposed algorithm.
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This technical report presents a variational Bayes algorithm for semisupervised hyperspectral image unmixing. The presented Bayesian model employs a heavy tailed, nonnegatively truncated Laplace prior over the abundance coefficients. This prior imposes both the sparsity assumption and the nonnegativity constraint on the abundance coefficients. Experimental results conducted on the Aviris Cuprite data set are presented that demonstrate the effectiveness of the proposed method.
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents an online version of the widely used sparse Bayesian learning (SBL) algorithm. Exploiting the variational Bayes framework, an efficient online SBL algorithm is constructed, that acts as a fully automatic learning method for the adaptive estimation of sparse time-varying signals. The new method is based on second order statistics and comprises a simple, automated sparsity-imposing mechanism, different from that of other known schemes. The effectiveness of the proposed online Bayesian algorithm is illustrated using experimental results conducted on synthetic data. These results show that the proposed scheme achieves faster initial convergence and superior estimation performance compared to other related state-of-the-art schemes.
    ICASSP 2014 - 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP); 05/2014
  • Spyridoula D. Xenaki, Konstantinos D. Koutroumbas, Athanasios A. Rontogiannis
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper a new sparse adaptive possibilistic clustering algorithm is presented. The algorithm exhibits high immunity to outliers and provides improved estimates of the cluster representatives by adjusting dynamically certain critical parameters. In addition, the proposed scheme manages - in principle - to estimate the actual number of clusters and by properly imposing sparsity, it becomes capable to deal well with closely located clusters of different densities. Extensive experimental results verify the previous statements.
    ICASSP 2014 - 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP); 05/2014
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Recently, a number of mostly $\ell_1$-norm regularized least squares type deterministic algorithms have been proposed to address the problem of \emph{sparse} adaptive signal estimation and system identification. From a Bayesian perspective, this task is equivalent to maximum a posteriori probability estimation under a sparsity promoting heavy-tailed prior for the parameters of interest. Following a different approach, this paper develops a unifying framework of sparse \emph{variational Bayes} algorithms that employ heavy-tailed priors in conjugate hierarchical form to facilitate posterior inference. The resulting fully automated variational schemes are first presented in a batch iterative form. Then it is shown that by properly exploiting the structure of the batch estimation task, new sparse adaptive variational Bayes algorithms can be derived, which have the ability to impose and track sparsity during real-time processing in a time-varying environment. The most important feature of the proposed algorithms is that they completely eliminate the need for computationally costly parameter fine-tuning, a necessary ingredient of sparse adaptive deterministic algorithms. Extensive simulation results are provided to demonstrate the effectiveness of the new sparse variational Bayes algorithms against state-of-the-art deterministic techniques for adaptive channel estimation. The results show that the proposed algorithms are numerically robust and exhibit in general superior estimation performance compared to their deterministic counterparts.
    IEEE Transactions on Signal Processing 01/2014; 62(18). DOI:10.1109/TSP.2014.2338839 · 3.20 Impact Factor
  • IGARSS; 01/2014
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A non trivial problem that arises in several applications is the estimation of the mean of a truncated normal distribution. In this paper, an iterative deterministic scheme for approximating this mean is proposed. It has been inspired from an iterative Markov chain Monte Carlo (MCMC) scheme that addresses this problem and it can be viewed as a generalization of a recently proposed relevant model. Conditions are provided under which it is proved that the scheme converges to a unique fixed point. Finally, the theoretical results are also supported by computer simulations, which also show the rapid convergence of the method to a solution vector that is very close to the mean of the truncated normal distribution under study.
  • IEEE WHISPERS: 5th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing; 06/2013
  • Konstantinos D Koutroumbas
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, a new methodology for curve approximation is presented. The method is suitable for both self-intersected and non self-intersected curves, it combines elements from graph theory and from elementary geometry and it is fully automated. More specifically, graph theory tools are utilized in order: 1) to remove the details that are irrelevant to the general shape of the curve under study and 2) to decompose the curve into non self-intersecting smaller curves. Then, each such smaller curve is processed via geometrical tools in order to approximate it efficiently with linear segments. Experimental results show that the proposed method compares well with several other methods of the same purpose.
    IEEE Transactions on Image Processing 05/2012; 21(9):3877-87. DOI:10.1109/TIP.2012.2199507 · 3.11 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper the problem of semisupervised hyper spectral unmixing is considered. More specifically, the unmixing process is formulated as a linear regression problem, where the abundance's physical constraints are taken into account. Based on this formulation, a novel hierarchical Bayesian model is proposed and suitable priors are selected for the model parameters such that, on the one hand, they ensure the nonnegativity of the abundances, while on the other hand they favor sparse solutions for the abundances' vector. Performing Bayesian inference based on the proposed hierarchical Bayesian model, a new low-complexity iterative method is derived, and its connection with Gibbs sampling and variational Bayesian inference is highlighted. Experimental results on both synthetic and real hyperspectral data illustrate that the proposed method converges fast, favors sparsity in the abun dances' vector, and offers improved estimation accuracy compared to other related methods.
    IEEE Transactions on Signal Processing 02/2012; 60(2):585-599. DOI:10.1109/TSP.2011.2174052 · 3.20 Impact Factor
  • A.A. Rontogiannis, K.E. Themelis, K.D. Koutroumbas
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a novel hierarchical Bayesian model which allows to reconstruct sparse signals using a set of linear measurements corrupted by Gaussian noise. The proposed model can be considered as the Bayesian counterpart of the adaptive lasso criterion. A fast iterative algorithm, which is based on the type-II maximum likelihood methodology, is properly adjusted to conduct Bayesian inference on the unknown model parameters. The performance of the proposed hierarchical Bayesian approach is illustrated on the reconstruction of both sparse synthetic data, as well as real images. Experimental results show the improved performance of the proposed approach, when compared to state-of-the-art Bayesian compressive sensing algorithms.
    Signal Processing Conference (EUSIPCO), 2012 Proceedings of the 20th European; 01/2012
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This article presents a comparative study of three different types of estimators used for supervised linear unmixing of two MEx/OMEGA hyperspectral cubes. The algorithms take into account the constraints of the abundance fractions, in order to get physically interpretable results. Abundance maps show that the Bayesian maximum a posteriori probability (MAP) estimator proposed in Themelis and Rontogiannis (2008) outperforms the other two schemes, offering a compromise between complexity and estimation performance. Thus, the MAP estimator is a candidate algorithm to perform ice and minerals detection on large hyperspectral datasets.
    Planetary and Space Science 12/2011; 68(1). DOI:10.1016/j.pss.2011.11.015 · 1.63 Impact Factor
  • K. Koutroumbas, I. Tsagouri, A. Belehaki
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents an analysis of a set of time series that represent foF2 disturbances during storm conditions, using clustering tools. The time series under study have been drawn from ionospheric observations obtained from eight European middle latitude ionosondes during a significant number of storm-time intervals and they are divided into eight groups according to the latitude (middle to low and middle to high) and the local time of the observation point at storm onset (afternoon, evening, morning, prenoon). The time series in each group have been analyzed using clustering-based methods. Specifically, each time series has been represented using two different ways of representation: the first is the raw representation while the second is through the parameters of the autoregressive (AR) model that best represents it. For each representation a hierarchy of clusterings is produced via the complete link algorithm. The two produced hierarchies are combined to a single one and the final clustering results are extracted from the produced hierarchy. The obtained results are in close agreement with the theoretical formulations concerning ionospheric storm effects at middle latitudes. In addition, they may be proved useful in the development of more accurate ionospheric forecasting methods.
    Advances in Space Research 05/2010; 45(9):1129-1144. DOI:10.1016/j.asr.2010.01.016 · 1.24 Impact Factor
  • 1st 04/2010; Academic Press (imprint of Elsevier Science)., ISBN: 978-0-12-374486-9
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this study two novel approaches for supervised and semi-supervised hyperspectral unmixing are applied in the unmixing of CHRIS/PROBA data, in order to monitor seasonal land cover changes - in particular plant foliar coverage. Foliar coverage variations are directly linked to seasonal changes of the ecophysiological status of a plant (i.e. growth status, pigment concentrations, LAI etc). The high potential of using hyperspectral satellite data in monitoring plant biochemical and structural characteristics is important in ecophysiological studies. A reliable and efficient method to extract leaf and/or canopy information from a mixed pixel significantly contributes towards this direction. In this study, the development of two efficient algorithms in spectral unmixing enables the detection and mapping of leaf contribution to the overall pixel spectra and its seasonal variations. For this purpose, leaf spectra measured in the field, simultaneously to satellite acquisitions, are included in the endmember data set. The proposed unmixing techniques are performed on ground reflectances, assuming knowledge of the number and spectral signatures of the objects present in the images. An efficient estimation for their corresponding fractions in the pixels of the image is developed, based on a recently proposed maximum a posteriori probability (MAP) method. By exploiting the constraints naturally imposed to the problem, closed form expressions are derived for the statistical parameters required by the MAP estimator. In the semi-supervised scenario, we assume that a spectral library is given, containing spectral signatures of multiple endmembers. The objective in the latter case, is (a) to determine how many and which endmembers are present in the mixed pixel under study and (b) to use the selected endmembers to estimate the corresponding abundance fractions - especially the abundance of foliar coverage. The approach is based on a properly modified weighted l1-regularized least squares algorithm. The motivation of using the sparsity promoting l1 norm is that in practice, only a small number of the available endmembers are present in each pixel. Based on this observation, a weighted version of the well-known least absolute shrinkage and selection operator (lasso) criterion is utilized, where weights are used for penalizing different coefficients in the l1-regularization scheme. To efficiently solve the l1 minimization problem, the Least Angle Regression (LARS) algorithm is used. The performance of the proposed methods is compared to the performance of existing unmixing algorithms, such as standard convex programming procedures and ENVI software unmixing routines.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Visible and infrared imaging spectroscopy is a key remote sensing technique used to study and monitor planets. It allows the detection, mapping and characterization of minerals, as well as volatile species, whose presence often provide clues for the resolution of Martian key climatic and geological issues. The OMEGA imaging spectrometer onboard Mars Express comprises 352 spectral channels from 0.3 to 5.1 μm at a spatial resolution ranging from 300 m to 4 km and provides the opportunity to investigate the mineralogy of the very top surface of Mars by looking at diagnostic spectral features in the visible and near infrared domains. The subject of this paper is to present a method of supervised unmixing applied on an OMEGA hyperspectral data cube. Data were obtained by looking to the South Polar Cap of Mars during local summer, where CO2 ice, water ice and dust were previously detected. This kind of prior knowledge on the above constituents (image endmembers) was obtained by using an unsupervised approach for hyperspectral data unmixing, namely the recently proposed Bayesian positive source separation (BSS) with positivity constraints [1]. We propose an efficient estimation for the endmembers' corresponding fractions in the pixels of the image, based on a recently proposed maximum a posteriori probability (MAP) method. This method is properly adjusted for hyperspectral data processing, and in particular for abundance estimation problems. To ensure the physical interpretation of the estimation results, the necessary properties of positivity and full additivity of the abundance fractions are both imposed. Closed form expressions are then derived for the statistical parameters required by the MAP estimator. Experimental results consisting of images of estimated abundance fractions are presented and compared to the corresponding existing results of the BSS algorithm. Simulated data are also used to demonstrate the performance and reliability of the proposed method. The proposed method offers significant computational savings as compared to Monte Carlo simulations or quadratic programming methods. This makes the algorithm suitable for (near) real-time processing of large hyperspectral data volumes including significant number of endmembers. First results show that the proposed approach can be a useful tool for providing efficient processing of hyperspectral data. [1] S. Moussaoui et al, "On the decomposition of Mars hyperspectral data by ICA and Bayesian positive source separation", Neurocomputing, vol. 71, pp. 2194-2208, June 2008.
  • Source
    Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2010, 14-19 March 2010, Sheraton Dallas Hotel, Dallas, Texas, USA; 01/2010
  • I. Tsagouri, K. Koutroumbas, A. Belehaki
    [Show abstract] [Hide abstract]
    ABSTRACT: The application of the SWIF (Solar Wind driven autoregression model for Ionospheric short term Forecast) model to the European Digital Upper Atmosphere Server - DIAS is discussed in this paper. SWIF is an empirical method designed to provide alerts and warnings for upcoming ionospheric storm disturbances as well as ionospheric forecasts for the middle latitude ionosphere. It combines historical and real-time ionospheric observations with solar wind parameters obtained in real time at L1 point through the cooperation of an autoregression forecasting algorithm, namely TSAR with an empirical ionospheric storm time model, namely STIM that is triggered by solar wind disturbances. SWIF can be considered as a powerful tool for the development of a full set of ionospheric forecasting services and therefore it can efficiently serve the objectives of the DIAS system (http://dias.space.noa.gr). As a first step towards the application of SWIF in DIAS, the work presented here includes the evaluation of SWIF's performance over all DIAS station locations (Athens, Rome, Ebre, Arenosillo, Chilton, Juliusruh, and Pruhonice) and the determination of its prediction accuracy limits for all possible geophysical conditions based on historical data. The results verify the reliability of SWIF's performance for prediction step up to 24 h ahead for all European locations and drive the upgrade of the SWIF algorithm for the accommodation of the prediction error at each prediction step. Finally, the on-line implementation of the SWIF algorithm in DIAS system leads to the substantial upgrade of all DIAS related products and services (e.g. single station forecasts, European forecasting maps, alerts and warnings). This upgrade will strengthen further the basis of the EURIPOS initiative, since reliable ionospheric forecasts are important for the development of new models and tools towards the successful specification of the ionosphere-plasmasphere system.
  • I. Tsagouri, K. Koutroumbas, A. Belehaki
    [Show abstract] [Hide abstract]
    ABSTRACT: The development of a new ionospheric forecasting algorithm, called the Solar Wind driven autoregression model for Ionospheric short-term Forecast (SWIF) is presented. SWIF combines historical and real-time ionospheric observations with solar wind parameters obtained in real time at the L1 point. This is achieved through the cooperation of an autoregression forecasting algorithm, called Time Series AutoRegressive (TSAR), with the empirical Storm Time Ionospheric Model that formulates the ionospheric storm time response based on solar wind input. The evaluation of SWIF's predictions was principally focused on its performance during selected storm time intervals over Europe. The results demonstrate significant improvement of SWIF's prediction capability over its predecessor, TSAR, significant improvement over climatology and evidence of SWIF's efficiency compared to other forecasting methods. Moreover, the evaluation of SWIF's output over long time periods, that include a wide range of geophysical conditions, suggests that SWIF can be used for prediction up to 24 h ahead.
    Radio Science 02/2009; 44(1). DOI:10.1029/2008RS004112 · 1.45 Impact Factor