A scalable algorithm is derived for multilevel quantization of sensor observations in distributed sensor networks, which consist of a number of sensors transmitting a summary information of their observations to the fusion center for a final decision. The proposed algorithm is directly minimizing the overall error probability of the network without resorting to minimizing pseudo objective functions such as distances between probability distributions. The problem formulation makes it possible to consider globally optimum error minimization at the fusion center and a person-by-person optimum quantization at each sensor. The complexity of the algorithm is quasi-linear for i.i.d. sensors. Experimental results indicate that the proposed scheme is superior in comparison to the current state-of-the-art.

Decentralized detection is studied for parallel-access sensor networks, where sensor statistics are not known completely and are assumed to follow distribution functions which belong to known uncertainty classes. It is shown that there exist no minimax robust tests over the deterministic decision rules for the uncertainty classes built with respect to the Kullback-Leibler (KL)-divergence. For the KL-divergence as well as for some other uncertainty classes, such as the α-divergences, the joint stochastic boundedness property, which is the fundamental rule to prove minimax robustness, fails to hold. This raises a natural question whether a solution to minimax robust decentralized detection problem can be given if the uncertainty classes do not own this property. An answer to this question has been shown to be positive, which leads to a generalization of an existing work. Moreover, it is shown that for Huber's extended uncertainty classes quantization functions at the sensors are not required to be monotone in order to claim minimax robustness. A possible generalization of the theory to minimax- and Neyman-Pearson formulations, repeated observations, imperfect reporting channels and different network topologies have been discussed. Simulation examples are provided considering clipped- and censored likelihood ratio tests.

Minimax robust hypothesis testing is studied for the cases where the collected data samples are corrupted by outliers and are mismodeled due to modeling errors. For the former case, Huber's clipped likelihood ratio test is introduced and analyzed. For the latter case, first, a robust hypothesis testing scheme based on the Kullback-Leibler divergence is designed. This approach generalizes a previous work by Levy. Second, Dabak and Johnson's asymptotically robust test is introduced, and other possible designs based on f-divergences are investigated. All proposed and analyzed robust tests are extended to fixed sample size and sequential probability ratio tests. Simulations are provided to exemplify and evaluate the theoretical derivations.

We propose a minimax robust hypothesis testing strategy between two composite hypotheses determined by the neighborhoods of two nominal distributions with respect to the squared Hellinger distance. The robust tests obtained are the nonlinearly transformed versions of the nominal likelihood ratios, whereas the least favorable densities are derived in three different regions. In two of them, they are scaled versions of the corresponding nominal densities and in the third region they form a composite version of the two nominal densities. The outcomes and implications of the proposed robust test are discussed through comparisons with the recent literature.

Blind steganalysis of JPEG images is addressed by modeling the correlations among the DCT coefficients usingK-variate (K≥ 2) p.d.f. estimates (p.d.f.s) constructed by means of Markov random field (MRF) cliques. The reasoning of using high variate p.d.f.s together with MRF cliques for image steganalysis is explained via a classical detection problem. Although our approach has many improvements over the current state-of-the-art, it suffers from the high dimensionality and the sparseness of the high variate p.d.f.s. The dimensionality problem as well as the sparseness problem are solved heuristically by means of dimensionality reduction and feature selection algorithms. The detection accuracy of the proposed method(s) is evaluated over Memon's (30.000 images) and Goljan's (1912 images) image sets. It is shown that practically applicable steganalysis systems are possible with a suitable dimensionality reduction technique and these systems can provide, in general, improved detection accuracy over the current state-of-the-art. Experimental results also justify this assertion.

In order to reduce the computational complexity of maximum-likelihood symbol estimation (MLSE) demapping of multilevel codes which is based on block partitioning and which produces soft input for a multistage decoding (MSD) process, three different demapping algorithms are proposed. It is theoretically proven that the proposed algorithms can reduce exponentially increasing computational complexity of the MLSE demapping algorithm to a constant complexity (neglecting comparisons). It is shown by extensive simulations for AWGN and Rayleigh fading channels that the proposed low complexity demapping algorithms can achieve near MLSE performance.

We introduce the use of multidimensional (MD) constellations in a multilevel coding (MLC) scheme with multi-stage decoding (MSD) designed for broadcasting communications, where services with different quality of service (QoS) are desirable. We show that the number of different protection levels increases when using MD constellations. Besides, the appropriate block labeling (BL) partitioning for an MD constellation is found by applying the binary switching algorithm (BSA) as an efficient search algorithm. Two cost functions for the BSA are proposed based on the BL criterion. Some methods to construct an appropriate MD constellation are presented, including the use of uniform and non-uniform component constellations. The viability of the proposed MD constellations approach for broadcasting with different protection levels is evaluated analyzing the mutual information (MI) of each level. Finally, we present a comparison between a unidimensional and a multidimensional scheme which again demonstrates the benefits of the proposed scheme.

In this paper we present a novel method based on singular value decomposition (SVD) for forensic analysis of digital images. We show that image tampering distorts linear dependencies of image rows/columns and derived features can be accurate enough to detect image manipulations and digital forgeries. Extensive experiments show that the proposed approach can outperform the counterparts in the literature.

This paper is concerned with the universal (blind) image steganalysis problem and introduces a novel method to detect especially spatial domain steganographic methods. The proposed steganalyzer models linear dependencies of image rows/columns in local neighborhoods using singular value decomposition transform and employs content independency provided by a Wiener filtering process. Experimental results show that the novel method has superior performance when compared with its counterparts in terms of spatial domain steganography. Experiments also demonstrate the reasonable ability of the method to detect discrete cosine transform-based steganography as well as the perturbation quantization method.

This paper introduces a novel universal steganalysis method in order to attack especially spatial domain steganographic algorithms. The basic idea behind the proposed method is twofold: The first is the capability of modeling linear dependencies of image rows/columns in local neighborhoods and the second is the normalization of different energy levels for different images. This idea is simply realized with integrating the logarithms of singular values calculated over image sub blocks resulting to a novel staganalyser, named as "LogSv". We consider three popular spatial domain steganographic algorithms; LSB, Hide4pgp, Steghide, five DCT domain steganographic algorithms; F5, Outguess, JP Hide&Seek, MB1, MB2 and PQ. Performance of the proposed method is compared with the state-of-the-art steganalysers.

This paper presents a singular value decomposition (SVD) based steganalysis technique to determine the watermarking method used to embed a watermark in an image. The detection is carried out in three steps. First, the proposed technique determines whether an image under consideration contains a watermark. If a watermark is detected, the embedding domain is revealed. Finally, the exact watermarking algorithm is named. The idea behind the method is that when the image is watermarked, relative and strict linear dependencies of rows/columns will differ from the original image and this can be modeled by the analysis of SVD. By using SVD, several features for the classification of the original and watermarked images are defined. The classification operation including both a linear and a SVM classifier is performed with a feature selection algorithm, which serves to reduce the number of features and to increase the detection performance. The performance of the proposed technique is promising and simulation results indicate that the chosen features can reliably detect the watermarking domain as well as the watermarking method.

Perturbed quantization (PQ) data hiding is almost undetectable with the current steganalysis methods. We briefly describe PQ and propose singular value decomposition (SVD)-based features for the steganalysis of JPEG-based PQ data hiding in images. We show that JPEG-based PQ data hiding distorts linear dependencies of rows/columns of pixel values, and proposed features can be exploited within a simple classifier for the steganalysis of PQ. The proposed steganalyzer detects PQ embedding on relatively smooth stego images with 70% detection accuracy on average for different embedding rates.