By Topic

Signal Processing, IEEE Transactions on

Issue 12 • Date June15, 2014

Filter Results

Displaying Results 1 - 25 of 30
  • [Front cover]

    Publication Year: 2014 , Page(s): C1
    Save to Project icon | Request Permissions | PDF file iconPDF (208 KB)  
    Freely Available from IEEE
  • IEEE Transactions on Signal Processing publication information

    Publication Year: 2014 , Page(s): C2
    Save to Project icon | Request Permissions | PDF file iconPDF (135 KB)  
    Freely Available from IEEE
  • Table of contents

    Publication Year: 2014 , Page(s): 3025 - 3026
    Save to Project icon | Request Permissions | PDF file iconPDF (228 KB)  
    Freely Available from IEEE
  • Table of contents

    Publication Year: 2014 , Page(s): 3027 - 3028
    Save to Project icon | Request Permissions | PDF file iconPDF (229 KB)  
    Freely Available from IEEE
  • Consensus-based Distributed Particle Filtering With Distributed Proposal Adaptation

    Publication Year: 2014 , Page(s): 3029 - 3041
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2954 KB) |  | HTML iconHTML  

    We develop a distributed particle filter for sequential estimation of a global state in a decentralized wireless sensor network. A global state estimate that takes into account the measurements of all sensors is computed in a distributed manner, using only local calculations at the individual sensors and local communication between neighboring sensors. The paper presents two main contributions. First, the likelihood consensus scheme for distributed calculation of the joint likelihood function (used by the local particle filters) is generalized to arbitrary local likelihood functions. This generalization overcomes the restriction to exponential-family likelihood functions that limited the applicability of the original likelihood consensus (Hlinka et al., “Likelihood consensus and its application to distributed particle filtering,” IEEE Trans. Signal Process., vol. 60, pp. 4334-4349, Aug. 2012). The second contribution is a consensus-based distributed method for adapting the proposal densities used by the local particle filters. This adaptation takes into account the measurements of all sensors, and it can yield a significant performance improvement or, alternatively, a significant reduction of the number of particles required for a given level of accuracy. The performance of the proposed distributed particle filter is demonstrated for a target tracking problem. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Discrete Signal Processing on Graphs: Frequency Analysis

    Publication Year: 2014 , Page(s): 3042 - 3054
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2094 KB) |  | HTML iconHTML  

    Signals and datasets that arise in physical and engineering applications, as well as social, genetics, biomolecular, and many other domains, are becoming increasingly larger and more complex. In contrast to traditional time and image signals, data in these domains are supported by arbitrary graphs. Signal processing on graphs extends concepts and techniques from traditional signal processing to data indexed by generic graphs. This paper studies the concepts of low and high frequencies on graphs, and low-, high- and band-pass graph signals and graph filters. In traditional signal processing, these concepts are easily defined because of a natural frequency ordering that has a physical interpretation. For signals residing on graphs, in general, there is no obvious frequency ordering. We propose a definition of total variation for graph signals that naturally leads to a frequency ordering on graphs and defines low-, high-, and band-pass graph signals and filters. We study the design of graph filters with specified frequency response, and illustrate our approach with applications to sensor malfunction detection and data classification. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optimal Periodic Sensor Scheduling in Networks of Dynamical Systems

    Publication Year: 2014 , Page(s): 3055 - 3068
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3239 KB) |  | HTML iconHTML  

    We consider the problem of finding optimal time-periodic sensor schedules for estimating the state of discrete-time dynamical systems. We assume that multiple sensors have been deployed and that the sensors are subject to resource constraints, which limits the number of times each can be activated over one period of the periodic schedule. We seek an algorithm that strikes a balance between estimation accuracy and total sensor activations over one period. We make a correspondence between active sensors and the nonzero columns of the estimator gain. We formulate an optimization problem in which we minimize the trace of the error covariance with respect to the estimator gain while simultaneously penalizing the number of nonzero columns of the estimator gain. This optimization problem is combinatorial in nature, and we employ the alternating direction method of multipliers (ADMM) to find its locally optimal solutions. Numerical results and comparisons with other sensor scheduling algorithms in the literature are provided to illustrate the effectiveness of our proposed method. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Collective Ratings for Online Communities With Strategic Users

    Publication Year: 2014 , Page(s): 3069 - 3083
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2930 KB) |  | HTML iconHTML  

    Despite the success of emerging online communities, they face a serious practical challenge: the participating agents are strategic, and incentive mechanisms are needed to compel such agents to provide high-quality services. Traditional mechanisms based on pricing and direct reciprocity schemes are not effective in providing incentives in such communities due to their unique features: large number of agents able to perform diverse services, imperfect monitoring of agents' service quality, etc. To compel agents to provide high-quality services, we develop a novel game-theoretic framework for providing incentives using rating-based pricing schemes. In our framework, the service-providing agents are not rated individually; instead, they are divided into separate groups based on their expertise, location, etc., and are rated collectively, as a group. A collective rating is updated for each group based on the quality of service provided by all the agents appertaining to the group. Depending on whether a group of agents collectively contributes a sufficiently high level of services or not, the agents in the group are rewarded or punished through increased or decreased collective rating, which will lead to higher or lower payments they receive in the future. We systematically analyze how the group size and the rating scheme affect the community designer's revenue as well as the social welfare of the agents and, based on this analysis. We design optimal rating protocols and show that these protocols can significantly improve the social welfare of the community as compared to a variety of alternative incentive mechanisms. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • SUMIS: Near-Optimal Soft-In Soft-Out MIMO Detection With Low and Fixed Complexity

    Publication Year: 2014 , Page(s): 3084 - 3097
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3373 KB) |  | HTML iconHTML  

    The fundamental problem of interest here is soft-input-soft-output multiple-input-multiple-output (MIMO) detection. We propose a method, referred to as subspace marginalization with interference suppression (SUMIS), that yields unprecedented performance at low and fixed (deterministic) complexity. Our method provides a well-defined tradeoff between computational complexity and performance. Apart from an initial sorting step consisting of selecting channel-matrix columns, the algorithm involves no searching nor algorithmic branching; hence the algorithm has a completely predictable run-time and allows for a highly parallel implementation. We numerically assess the performance of SUMIS in different practical settings: full/partial channel state information, sequential/iterative decoding, and low/high rate outer codes. We also comment on how the SUMIS method performs in systems with a large number of transmit antennas. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Distributed State Estimation With Dimension Reduction Preprocessing

    Publication Year: 2014 , Page(s): 3098 - 3110
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2981 KB) |  | HTML iconHTML  

    System state estimation relies heavily on the measurements. With the advance of sensing technology, the ability to measure is no longer a bottleneck in many systems, and more and more researchers now focus on the rich-information setting, i.e., big data. However, although information never hurts, it does not help unconditionally. How to make the most of it depends on whether we can process the data efficiently. In some systems, the inherent constraint such as the bandwidth and cost makes it necessary to reduce the dimension of the measurement before further processing. The problem that the raw measurements are first preprocessed to reduce size and then used for estimation is addressed in this paper. It is shown that there is a lower bound on the size of the preprocessed data such that if the size is beyond the bound, there exists a closed-form estimator design that the linear minimum mean-square estimation can be obtained. Moreover, we propose an algorithm that is guaranteed to converge to a stationary point to design an estimator in the conditions that the lower bound cannot be reached. Besides convergence, the proposed algorithm guarantees bounded performance loss compared with the global optimal solution under some additional conditions. Finally, simulation results in three different applications are shown to demonstrate the effectiveness of the proposed algorithm. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Data-Aided Channel Estimation in Large Antenna Systems

    Publication Year: 2014 , Page(s): 3111 - 3124
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3683 KB) |  | HTML iconHTML  

    This paper is concerned with a uplink scheme for multicell large antenna systems. We study a channel estimation technique where partially decoded data is used to estimate the channel. We show that there are two types of interference components in this scheme that do not vanish even when the number of antennas grows to infinity. The first type, referred to as cross-contamination, is due to the correlation among the data signals from different users. The second type, referred to as self-contamination, is due to the dependency between the channel estimate and the estimation error. Cross contamination is in principle similar to pilot contamination in a conventional pilot-based channel estimation scheme, while self-contamination is unique for the data-aided scheme. For efficient use of the channel, the data part in a signaling frame is typically much longer than the pilot part for a practical system. Consequently, compared with pilot signals, data signals naturally have lower cross correlation. This fact reduces the cross-contamination effect in the data-aided scheme. Furthermore, self-contamination can be effectively suppressed by iterative processing. These results are confirmed by both analyses and simulations. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Direct Construction of Superoscillations

    Publication Year: 2014 , Page(s): 3125 - 3134
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1810 KB) |  | HTML iconHTML  

    Oscillations of a bandlimited signal at a rate faster than its maximum frequency are called “superoscillations” and have been found useful e.g., in connection with superresolution and superdirectivity. We consider signals of fixed bandwidth and with a finite or infinite number of samples at the Nyquist rate, which are regarded as the adjustable signal parameters. We show that this class of signals can be made to superoscillate by prescribing its values on an arbitrarily fine and possibly nonuniform grid. The superoscillations can be made to occur at a large distance from the nonzero samples of the signal. We give necessary and sufficient conditions for the problem to have a solution, in terms of the nature of the two sets involved in the problem. Since the number of constraints can in general be different from the number of signal parameters, the problem can be exactly determined, underdetermined or overdetermined. We describe the solutions in each of these situations. The connection with oversampling and variational formulations is also discussed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Geolocation by Direction of Arrival Using Arrays With Unknown Orientation

    Publication Year: 2014 , Page(s): 3135 - 3142
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2010 KB) |  | HTML iconHTML  

    It takes a great deal of care to accurately align direction of arrival (DOA) sensors to a reference direction. Any error in alignment degrades the localization accuracy of the entire system. We propose a method that enables the alignment using DOA measurements of arbitrary sources. This saves the efforts to set and maintain the alignment by external means. To simplify the exhibition, it is assumed that the sensors and the sources are confined to a plane. The method is based on the maximum-likelihood estimator. The main challenge is the discontinuities in the cost function due to the circular nature of angle measurements. The proposed method is verified by simulations, and the performance is compared to lower bounds and also to systems with perfect alignment. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fixed-Lag Smoothing for Bayes Optimal Knowledge Exploitation in Target Tracking

    Publication Year: 2014 , Page(s): 3143 - 3152
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2871 KB) |  | HTML iconHTML  

    In this work, we are interested in the improvements attainable when multiscan processing of external knowledge is performed over a moving time window. We propose a novel algorithm that enforces the state constraints by using a Fixed-Lag Smoothing procedure within the prediction step of the Bayesian recursion. For proving the improvements, we utilize differential entropy as a measure of uncertainty and show that the approach guarantees a lower or equal posterior differential entropy than classical single-step constrained filtering. Simulation results using examples for single-target tracking are presented to verify that a Sequential Monte Carlo implementation of the proposed algorithm guarantees an improved tracking accuracy. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On Optimal Gaussian Signaling in MIMO Relay Channels With Partial Decode-and-Forward

    Publication Year: 2014 , Page(s): 3153 - 3164
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2772 KB) |  | HTML iconHTML  

    For Gaussian multiple-input multiple-output (MIMO) relay channels with partial decode-and-forward, the optimal type of input distribution is still an open question in general. Recent research has revealed that in some other scenarios with unknown optimal input distributions (e.g., interference channels), improper (i.e., noncircular) Gaussian distributions can outperform proper (circular) Gaussian distributions. In this paper, we show that this is not the case for partial decode-and-forward in the Gaussian MIMO relay channel with Gaussian transmit signals, i.e., we show that a proper Gaussian input distribution is the optimal one among all Gaussian distributions. In order to prove this property, an innovation covariance matrix is introduced, and a decomposition is performed by considering the optimization over this matrix as an outer problem. A key point for showing optimality of proper signals then is a reformulation that reveals that one of the subproblems is equivalent to a sum rate maximization in a two-user MIMO broadcast channel under a sum covariance constraint, for which the optimality of proper signals can be shown. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An Efficient Partial-Sum Network Architecture for Semi-Parallel Polar Codes Decoder Implementation

    Publication Year: 2014 , Page(s): 3165 - 3179
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3413 KB) |  | HTML iconHTML  

    Polar codes have recently received a lot of attention because of their capacity-achieving performance and low encoding and decoding complexity. The performance of the successive cancellation decoder (SCD) of the polar codes highly depends on that of the partial-sum network (PSN) implementation. Hence, in this work, an efficient PSN architecture is proposed, based on the properties of polar codes. First, a new partial-sum updating algorithm and the corresponding PSN architecture are introduced which achieve a delay performance independent of the code length. Moreover, the area complexity is also reduced. Second, for a high-performance and area-efficient semi-parallel SCD implementation, a folded PSN architecture is presented to integrate seamlessly with the folded processing element architecture. This is achieved by using a novel folded decoding schedule. As a result, both the critical path delay and the area (excluding the memory for folding) of the semi-parallel SCD are approximately constant for a large range of code lengths. The proposed designs are implemented in both FPGA and ASIC and compared with the existing designs. Experimental result shows that for polar codes with large code length, the decoding throughput is improved by more than 1.05 times and the area is reduced by as much as 50.4%, compared with the state-of-the-art designs. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Sequential Distributed Detection in Energy-Constrained Wireless Sensor Networks

    Publication Year: 2014 , Page(s): 3180 - 3193
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3159 KB) |  | HTML iconHTML  

    The recently proposed sequential distributed detector based on level-triggered sampling operates as simple as the decision fusion techniques and at the same time performs as well as the data fusion techniques. Hence, it is well suited for resource-constrained wireless sensor networks. However, in practical cases where sensors observe discrete-time signals, the random overshoot above or below the sampling thresholds considerably degrades the performance of the considered detector. We propose, for systems with stringent energy constraints, a novel approach to tackle this problem by encoding the overshoot into the time delay between the sampling time and the transmission time. Specifically, each sensor computes the local log-likelihood ratio (LLR) and samples it using level-triggered sampling. Then, it transmits a single pulse to the fusion center (FC) after a transmission delay that is proportional to the overshoot, as in pulse position modulation (PPM). The FC, upon receiving a bit decodes the corresponding overshoot and recovers the transmitted LLR value. It then updates the approximate global LLR and compares it with two threshold to either make a decision or to continue the sequential process. We analyze the asymptotic average detection delay performance of the proposed scheme. We then apply the proposed sequential scheme to target detection in wireless sensor networks under the four Swerling fluctuating target models. It is seen that the proposed sequential distributed detector offers significant performance advantage over conventional decision fusion techniques. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Radar Target Profiling and Recognition Based on TSI-Optimized Compressive Sensing Kernel

    Publication Year: 2014 , Page(s): 3194 - 3207
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3084 KB) |  | HTML iconHTML  

    The design of wideband radar systems is often limited by existing analog-to-digital (A/D) converter technology. State-of-the-art A/D rates and high effective number of bits result in rapidly increasing cost and power consumption for the radar system. Therefore, it is useful to consider compressive sensing methods that enable reduced sampling rate, and in many applications, prior knowledge of signals of interest can be learned from training data and used to design better compressive measurement kernels. In this paper, we use a task-specific information-based approach to optimizing sensing kernels for high-resolution radar range profiling of man-made targets. We employ a Gaussian mixture (GM) model for the targets and use a Taylor series expansion of the logarithm of the GM probability distribution to enable a closed-form gradient of information with respect to the sensing kernel. The GM model admits nuisance parameters such as target pose angle and range translation. The gradient is then used in a gradient-based approach to search for the optimal sensing kernel. In numerical simulations, we compare the performance of the proposed sensing kernel design to random projections and to lower-bandwidth waveforms that can be sampled at the Nyquist rate. Simulation results demonstrate that the proposed technique for sensing kernel design can significantly improve performance. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Distributed Detection in Tree Topologies With Byzantines

    Publication Year: 2014 , Page(s): 3208 - 3219
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3219 KB) |  | HTML iconHTML  

    In this paper, we consider the problem of distributed detection in tree topologies in the presence of Byzantines. The expression for minimum attacking power required by the Byzantines to blind the fusion center (FC) is obtained. More specifically, we show that when more than a certain fraction of individual node decisions are falsified, the decision fusion scheme becomes completely incapable. We obtain closed-form expressions for the optimal attacking strategies that minimize the detection error exponent at the FC. We also look at the possible countermeasures from the FC's perspective to protect the network from these Byzantines. We formulate the robust topology design problem as a bi-level program and provide an efficient algorithm to solve it. We also provide some numerical results to gain insights into the solution. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Approximate Subspace-Based Iterative Adaptive Approach for Fast Two-Dimensional Spectral Estimation

    Publication Year: 2014 , Page(s): 3220 - 3231
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (4024 KB) |  | HTML iconHTML  

    In this paper, we devise a new approach for fast implementation of two-dimensional (2-D) iterative adaptive approach (IAA) using single or multiple snapshots. Our underlying idea is to apply the subspace methodology in this nonparametric technique by performing the IAA on the dominant singular vectors extracted from the singular value decomposition (SVD) or higher-order SVD of the multidimensional observations. In doing so, 2-D IAA is approximately realized by multiple steps of 1-D IAA, implying that computational attractiveness is achieved particularly for large data size, number of grid points and/or snapshot number. Algorithms based on matrix and tensor operations are developed, and their implementation complexities are analyzed. Computer simulations are also included to compare the proposed approach with the state-of-the-art techniques in terms of resolution probability, spectral estimation performance and computational requirement. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optimal Energy Allocation for Energy Harvesting Transmitters With Hybrid Energy Storage and Processing Cost

    Publication Year: 2014 , Page(s): 3232 - 3245
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3683 KB) |  | HTML iconHTML  

    We consider data transmission with an energy harvesting transmitter that has hybrid energy storage with a perfect super-capacitor (SC) and an inefficient battery. The SC has finite storage space while the battery has unlimited space. The transmitter can choose to store the harvested energy in the SC or in the battery. The energy is drained from the SC and the battery simultaneously. In this setting, we consider throughput optimal offline energy allocation problem over a point-to-point channel. In contrast to previous works, the hybrid energy storage model with finite and unlimited storage capacities imposes a generalized set of constraints on the transmission policy. As such, we show that the solution generalizes that for a single battery and is found by a sequential application of the directional water-filling algorithm. Next, we consider offline throughput maximization in the presence of an additive time-linear processing cost in the transmitter's circuitry. In this case, the transmitter has to additionally decide on the portions of the processing cost to be drained from the SC and the battery. Despite this additional complexity, we show that the solution is obtained by a sequential application of a directional glue pouring algorithm, parallel to the costless processing case. Finally, we provide numerical illustrations for optimal policies and performance comparisons with some heuristic online policies. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The Labeled Multi-Bernoulli Filter

    Publication Year: 2014 , Page(s): 3246 - 3260
    Cited by:  Papers (4)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (3698 KB) |  | HTML iconHTML  

    This paper proposes a generalization of the multi- Bernoulli filter called the labeled multi-Bernoulli filter that outputs target tracks. Moreover, the labeled multi-Bernoulli filter does not exhibit a cardinality bias due to a more accurate update approximation compared to the multi-Bernoulli filter by exploiting the conjugate prior form for labeled Random Finite Sets. The proposed filter can be interpreted as an efficient approximation of the δ-Generalized Labeled Multi-Bernoulli filter. It inherits the advantages of the multi-Bernoulli filter in regards to particle implementation and state estimation. It also inherits advantages of the δ-Generalized Labeled Multi-Bernoulli filter in that it outputs (labeled) target tracks and achieves better performance. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Distributed Compressive CSIT Estimation and Feedback for FDD Multi-User Massive MIMO Systems

    Publication Year: 2014 , Page(s): 3261 - 3271
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (2172 KB) |  | HTML iconHTML  

    To fully utilize the spatial multiplexing gains or array gains of massive MIMO, the channel state information must be obtained at the transmitter side (CSIT). However, conventional CSIT estimation approaches are not suitable for FDD massive MIMO systems because of the overwhelming training and feedback overhead. In this paper, we consider multi-user massive MIMO systems and deploy the compressive sensing (CS) technique to reduce the training as well as the feedback overhead in the CSIT estimation. The multi-user massive MIMO systems exhibits a hidden joint sparsity structure in the user channel matrices due to the shared local scatterers in the physical propagation environment. As such, instead of naively applying the conventional CS to the CSIT estimation, we propose a distributed compressive CSIT estimation scheme so that the compressed measurements are observed at the users locally, while the CSIT recovery is performed at the base station jointly. A joint orthogonal matching pursuit recovery algorithm is proposed to perform the CSIT recovery, with the capability of exploiting the hidden joint sparsity in the user channel matrices. We analyze the obtained CSIT quality in terms of the normalized mean absolute error, and through the closed-form expressions, we obtain simple insights into how the joint channel sparsity can be exploited to improve the CSIT recovery performance. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An Adaptive Approach to Learn Overcomplete Dictionaries With Efficient Numbers of Elements

    Publication Year: 2014 , Page(s): 3272 - 3283
    Multimedia
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (4297 KB) |  | HTML iconHTML  

    Dictionary learning for sparse representation has recently attracted attention among the signal processing society in a variety of applications such as denoising, classification, and compression. The number of elements in a learned dictionary is crucial since it governs specificity and optimality of sparse representation. Sparsity level, number of dictionary elements, and representation error are three correlated factors in which setting each pair of them results in a specific value of the third factor. An arbitrary selection of the number of dictionary elements affects either the sparsity level or/and the representation error. Despite recent advancements in training dictionaries, the number of dictionary elements is still heuristically set. To avoid the representation's suboptimality, a systematic approach to adapt the elements' number based on input datasets is essential. Some existing methods try to address this requirement such as enhanced K-SVD, sub-clustering K-SVD, and stage-wise K-SVD. However, it is not specified under which sparsity level and representation error criteria their learned dictionaries are size-optimized. We propose a new dictionary learning approach that automatically learns a dictionary with an efficient number of elements that provides both desired representation error and desired average sparsity level. In our proposed method, for any given representation error and average sparsity level, the number of elements in the learned dictionary varies based on content complexity of training datasets. The performance of the proposed method is demonstrated in image denoising. The proposed method is compared to state-of-the-art, and results confirm the superiority of the proposed approach. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • IEEE Transactions on Signal Processing EDICS

    Publication Year: 2014 , Page(s): 3284
    Save to Project icon | Request Permissions | PDF file iconPDF (101 KB)  
    Freely Available from IEEE

Aims & Scope

IEEE Transactions on Signal Processing covers novel theory, algorithms, performance analyses and applications of techniques for the processing, understanding, learning, retrieval, mining, and extraction of information from signals

Full Aims & Scope

Meet Our Editors

Editor-in-Chief
Sergios Theodoridis
University of Athens