Abbas Kazerooni, Azarang Golmohammadi, Farokh Marvasti
In this paper, we propose a new method for Salt-and-Pepper noise removal from images. Whereas most of the existing methods are based on Ordered Statistics filters, our method is based on the growing theory of Sparse Signal Processing. In other words, we convert the problem of denoising into a sparse signal reconstruction problem which can be dealt with the corresponding techniques. As a result, the output image of our method is preserved from the undesirable opacity which is a disadvantage of most of the other methods. We also introduce an efficient reconstruction algorithm which will be used in our method. Simulation results indicate that our method outperforms the other best-known methods both in term of PSNR and visual criterion. Furthermore, our method can be easily used for reconstruction of missing samples in erasure channels.
Amirhossein Javaheri, Jiaxi Ying, Daniel P. Palomar, Farokh Marvasti
Graph models provide efficient tools to capture the underlying structure of data defined over networks. Many real-world network topologies are subject to change over time. Learning to model the dynamic interactions between entities in such networks is known as time-varying graph learning. Current methodology for learning such models often lacks robustness to outliers in the data and fails to handle heavy-tailed distributions, a common feature in many real-world datasets (e.g., financial data). This paper addresses the problem of learning time-varying graph models capable of efficiently representing heavy-tailed data. Unlike traditional approaches, we incorporate graph structures with specific spectral properties to enhance data clustering in our model. Our proposed method, which can also deal with noise and missing values in the data, is based on a stochastic approach, where a non-negative vector auto-regressive (VAR) model captures the variations in the graph and a Student-t distribution models the signal originating from this underlying time-varying graph. We propose an iterative method to learn time-varying graph topologies within a semi-online framework where only a mini-batch of data is used to update the graph. Simulations with both synthetic and real datasets demonstrate the efficacy of our model in analyzing heavy-tailed data, particularly those found in financial markets.
Reza K. Farsani, Farokh Marvasti
In this paper, the Interference Network with General Message Sets (IN-GMS) is introduced in which several transmitters send messages to several receivers: Each subset of transmitters transmit an individual message to each subset of receivers. For such a general scenario, an achievability scheme is presented using the random coding. This scheme is systematically built based on the capacity achieving scheme for the Multiple Access Channel (MAC) with common message as well as the best known achievability scheme for the Broadcast Channel (BC) with common message. A graphical illustration of the random codebook construction procedure is also provided, by using which the achievability scheme is easily understood. Some benefits of the proposed achievability scheme are described. It is also shown that the resulting rate region is optimal for a class of orthogonal INs-GMS, which yields the capacity region. Finally, it is demonstrated that how this general achievability scheme can be used to derive capacity inner bounds for interference networks with different distribution of messages; in most cases, the proposed achievability scheme leads to the best known capacity inner bound for the underlying channel. Capacity inner bounds can also be derived for new communication scenarios.
Amirhossein Javaheri, Hadi Zayyani, Farokh Marvasti
This paper investigates the problem of recovering missing samples using methods based on sparse representation adapted especially for image signals. Instead of $l_2$-norm or Mean Square Error (MSE), a new perceptual quality measure is used as the similarity criterion between the original and the reconstructed images. The proposed criterion called Convex SIMilarity (CSIM) index is a modified version of the Structural SIMilarity (SSIM) index, which despite its predecessor, is convex and uni-modal. We derive mathematical properties for the proposed index and show how to optimally choose the parameters of the proposed criterion, investigating the Restricted Isometry (RIP) and error-sensitivity properties. We also propose an iterative sparse recovery method based on a constrained $l_1$-norm minimization problem, incorporating CSIM as the fidelity criterion. The resulting convex optimization problem is solved via an algorithm based on Alternating Direction Method of Multipliers (ADMM). Taking advantage of the convexity of the CSIM index, we also prove the convergence of the algorithm to the globally optimal solution of the proposed optimization problem, starting from any arbitrary point. Simulation results confirm the performance of the new similarity index as well as the proposed algorithm for missing sample recovery of image patch signals.
Nematollah Zarmehi, Farokh Marvasti
In this paper, we propose an Adaptive Singular Value Thresholding (ASVT) for low rank recovery under affine constraints. Unlike previous iterative methods that the threshold level is independent of the iteration number, in our proposed method, the threshold in adaptively decreases during iterations. The simulation results reveal that we get better performance with this thresholding strategy.
Hossein Hosseini, Farzad Hessar, Farokh Marvasti
In this paper, we propose a method for real-time high density impulse noise suppression from images. In our method, we first apply an impulse detector to identify the corrupted pixels and then employ an innovative weighted-average filter to restore them. The filter takes the nearest neighboring interpolated image as the initial image and computes the weights according to the relative positions of the corrupted and uncorrupted pixels. Experimental results show that the proposed method outperforms the best existing methods in both PSNR measure and visual quality and is quite suitable for real-time applications.
Ali Mousavi, Pedram Pad, Farokh Marvasti
Recently, a new class of binary codes for overloaded CDMA systems are proposed that not only has the ability of errorless communication but also suitable for detecting active users. These codes are called COWDA [1]. In [1], a Maximum Likelihood (ML) decoder is proposed for this class of codes. Although the proposed scheme of coding/decoding show impressive performance, the decoder can be improved. In this paper by assuming more practical conditions for the traffic in the system, we suggest an algorithm that increases the performance of the decoder several orders of magnitude (the Bit-Error-Rate (BER) is divided by a factor of 400 in some Eb/N0's The algorithm supposes the Poison distribution for the time of activation/deactivation of the users.
Mohammad Ali Sedaghat, Ralf Müller, Farokh Marvasti
We extend the result by Tse and Verdú on the optimum asymptotic multiuser efficiency of randomly spread CDMA with Binary Phase Shift Keying (BPSK) input. Random Gaussian and random binary antipodal spreading are considered. We obtain the optimum asymptotic multiuser efficiency of a $K$-user system with spreading gain $N$ when $K$ and $N\rightarrow\infty$ and the loading factor, $\frac{K}{N}$, grows logarithmically with $K$ under some conditions. It is shown that the optimum detector in a Gaussian randomly spread CDMA system has a performance close to the single user system at high Signal to Noise Ratio (SNR) when $K$ and $N\rightarrow\infty$ and the loading factor, $\frac{K}{N}$, is kept less than $\frac{\log_3K}{2}$. Random binary antipodal matrices are also studied and a lower bound for the optimum asymptotic multiuser efficiency is obtained. Furthermore, we investigate the connection between detecting matrices in the coin weighing problem and optimum asymptotic multiuser efficiency. We obtain a condition such that for any binary input, an $N\times K$ random matrix whose entries are chosen randomly from a finite set, is a detecting matrix as $K$ and $N\rightarrow \infty$.
Ahmadreza Moradipari, Sina Shahsavari, Ashkan Esmaeili, Farokh Marvasti
Inference and Estimation in Missing Information (MI) scenarios are important topics in Statistical Learning Theory and Machine Learning (ML). In ML literature, attempts have been made to enhance prediction through precise feature selection methods. In sparse linear models, LASSO is well-known in extracting the desired support of the signal and resisting against noisy systems. When sparse models are also suffering from MI, the sparse recovery and inference of the missing models are taken into account simultaneously. In this paper, we will introduce an approach which enjoys sparse regression and covariance matrix estimation to improve matrix completion accuracy, and as a result enhancing feature selection preciseness which leads to reduction in prediction Mean Squared Error (MSE). We will compare the effect of employing covariance matrix in enhancing estimation accuracy to the case it is not used in feature selection. Simulations show the improvement in the performance as compared to the case where the covariance matrix estimation is not used.
Mahdi Shamsi, Hadi Zayyani, Farokh Marvasti
Reconfigurable Intelligent Surfaces (RIS) have emerged as a transformative technology in wireless communication, enabling dynamic control over signal propagation. This paper tackles the challenge of mitigating Channel State Information (CSI) perturbations in RIS-aided systems, particularly for secure communication scenarios. Leveraging a first-order approximation technique, we develop a robust approach that strengthens the resilience of RIS configurations against CSI imperfections. The study considers both untrusted user interception and stealth radar applications, focusing on optimizing signal reflection and transmission in the presence of eavesdroppers. Simulation results demonstrate notable gains in security and efficiency while maintaining low computational complexity. By extending the stability range, the proposed method updates RIS elements using only a few matrix-vector multiplications, eliminating the need for repeated inverse or pseudo-inverse computations under small channel perturbations. Additionally, the framework provides a baseline for quantifying algorithmic sensitivity to CSI variations. Overall, the findings underscore the potential of RIS to enable secure and reliable communication in next-generation networks such as 6G.