Search

© 1991-2012 IEEE. Multiway datasets are widespread in signal processing and play an important role in blind signal separation, array processing, and biomedical signal processing, among others. One key strength of tensors is that their decompositions are unique under mild conditions, which allows the recovery of features or source signals. In several applications, such as classification, we wish to compare factor matrices of the decompositions. Though this is possible by first computing the tensor decompositions and subsequently comparing the factors, these decompositions are often computationally expensive. In this paper, we present a similarity method that indicates whether the factors in two modes are essentially equal without explicitly computing them. Essential equality conditions, which ensure the theoretical validity of our approach, are provided for various underlying tensor decompositions. The developed algorithm provides a computationally efficient way to compare factors. The method is illustrated in a context of emitter movement detection and fluorescence data analysis. ispartof: IEEE Transactions on Signal Processing vol:66 issue:5 pages:1273-1285 status: published

© 2017 IEEE. Nonnegative matrix factorizationisakey toolinmany data analysis applications such as feature extraction, compression, and noise filtering. Many existing algorithms impose additional constraintstotake into account prior knowledgeandtoimprove the physical interpretation. This letter proposes a novel algorithm for nonnegative matrix factorization, in which the factors are modeled by nonnegative polynomials. Using a parametric representation of finite-interval nonnegative polynomials, we obtain an optimization problem without external nonnegativity constraints, which can be solved using conventional quasi-Newton or nonlinear least-squares methods. The polynomial model guarantees smooth solutions and may realize a noise reduction. A dedicated orthogonal compression enables a significant reduction of the matrix dimensions, without sacrificing accuracy. The overall approach scales well to large matrices. The approach is illustrated with applications in hyperspectral imaging and chemical shift brain imaging. ispartof: IEEE Signal Processing Letters vol:24 issue:7 pages:948-952 status: published

© 1991-2012 IEEE. Many real-life signals are compressible, meaning that they depend on much fewer parameters than their sample size. In this paper, we use low-rank matrix or tensor representations for signal compression. We propose a new deterministic method for blind source separation that exploits the low-rank structure, enabling a unique separation of the source signals and providing a way to cope with large-scale data. We explain that our method reformulates the blind source separation problem as the computation of a tensor decomposition, after reshaping the observed data matrix into a tensor. This deterministic tensorization technique is called segmentation and is closely related to Hankel-based tensorization. We apply the same strategy to the mixing coefficients of the blind source separation problem, as in many large-scale applications, the mixture is also compressible because of many closely located sensors. Moreover, we combine both strategies, resulting in a general technique that allows us to exploit the underlying compactness of the sources and the mixture simultaneously. We illustrate the techniques for fetal electrocardiogram extraction and direction-of-arrival estimation in large-scale antenna arrays. ispartof: IEEE Transactions on Signal Processing vol:65 issue:2 pages:346-358 status: published

© 2016 IEEE. A new method for the blind identification of largescale finite impulse response (FIR) systems is presented. It exploits the fact that the system coefficients in large-scale problems often depend on much fewer parameters than the total number of entries in the coefficient vectors. We use low-rank models to compactly represent matricized versions of these compressible system coefficients.We show that blind system identification (BSI) then reduces to the computation of a structured tensor decomposition by using a deterministic tensorization technique called segmentation on the observed outputs. This careful exploitation of the low-rank structure enables the unique identification of both the system coefficients and the inputs. The approach does not require the input signals to be statistically independent. ispartof: pages:2015-2019 ispartof: Proc. 24th European Signal Processing Conference vol:2016-November pages:2015-2019 ispartof: EUSIPCO 2016 location:Budapest, Hungary date:Sep - Sep 2016 status: published

© 1991-2012 IEEE. A new blind signal separation (BSS) technique is proposed, enabling a deterministic separation of signals into rational functions. Rational functions can take on a wide range of forms, such as the well-known pole-like shape. The approach is a possible alternative for the well-known independent component analysis when the theoretical sources are not independent, such as for frequency spectra, or when only a small number of samples is available. The technique uses a low-rank decomposition on the tensorized version of the observed data matrix. The deterministic tensorization with Löwner matrices is comprehensively analyzed in this paper. Uniqueness properties are investigated, and a connection with the separation into exponential polynomials is made. Finally, the technique is illustrated for fetal electrocardiogram extraction and with an application in the domain of fluorescence spectroscopy, enabling the identification of chemical analytes using only a single excitation-emission matrix. ispartof: IEEE Transactions on Signal Processing vol:64 issue:8 pages:1909-1918 status: published

Higher-order tensors and their decompositions are abundantly present in domains such as signal processing (e.g., higher-order statistics [1] and sensor array processing [2]), scientific computing (e.g., discretized multivariate functions [3]?[6]), and quantum information theory (e.g., representation of quantum many-body states [7]). In many applications, the possibly huge tensors can be approximated well by compact multilinear models or decompositions. Tensor decompositions are more versatile tools than the linear models resulting from traditional matrix approaches. Compared to matrices, tensors have at least one extra dimension. The number of elements in a tensor increases exponentially with the number of dimensions, and so do the computational and memory requirements. The exponential dependency (and the problems that are caused by it) is called the curse of dimensionality. The curse limits the order of the tensors that can be handled. Even for a modest order, tensor problems are often large scale. Large tensors can be handled, and the curse can be alleviated or even removed by using a decomposition that represents the tensor instead of using the tensor itself. However, most decomposition algorithms require full tensors, which renders these algorithms infeasible for many data sets. If a tensor can be represented by a decomposition, this hypothesized structure can be exploited by using compressed sensing (CS) methods working on incomplete tensors, i.e., tensors with only a few known elements. © 2014 IEEE. ispartof: IEEE Signal Processing Magazine vol:31 issue:5 pages:71-79 status: published