مجله پردازش سیگنال پیشرفته
سال سوم شماره 2 (پیاپی 4، پاییز و زمستان 1398)

Breast cancer is the most common type of cancer that affects the female population. Early detection of cancer can increase the chance of treatment and is also the most effective way to fight the disease. The development of automated methods for the detection of cancer or tumor mass has been of interest to researchers. In this paper, a method based on deep convolutional neural networks for detecting tumor area from MRI images is introduced. The proposed method is to collect MRI images along with GT images from their tumor area and expand the data to train and test the neural network. The type of learning method used in this paper is supervised learning. The algorithm is tested on the RIDER breast dataset and the results show that the proposed method performs better than other image detection methods such as clustering methods. Benefits include high quality in tumor detection and acceptable speed at runtime.

In this paper, standardization of human brain MRI images is performed by applying several statistical methods The goal will be more clearly displayed after the standardization phase by removing the blurry and black noise spots, the border of the tumor areas and the different parts of the brain and cerebral fluid, which can be divided into three sections: white, gray and cerebral fluid. When a standard protocol is available, hybridization-based standardization and segmentation methods are one of the most appropriate tools for medical image segmentation. This method, despite its high accuracy, is time consuming and lengthy due to the high volume of computing. Methods based on statistical concepts are fashion, median, and mean, respectively, which are simple and explicit processing to apply to existing data from human brain MRI images. Boundary, decimal, and percentile methods are also used with simple concepts to extract milestones from existing data. To evaluate the performance of this paper, the scheme is first systematically simulated in MATLAB software, and then, for the area of comparison and comparison with the pre-standardization images, a software term Uvneti 8 was used to extract the gray layer. The results of the processing of this paper are visualized from standard images and statistical analysis of error and coefficient of variation are obtained and finally the most efficient method is extracted.

Extensive research has been conducted in the field of music signal processing which targeted context-based music information retrieval. Unfortunately, research on the computer-based processing of the traditional Persian music is rare, which is due to lack of standard databases. In this paper, a database, named Nava, is introduced for two basic tasks of the traditional Persian music field, Dastgah classification and instrument recognition. In terms of instrument, Dastgah and artist, Nava has enough comprehensiveness and variety. It contains the sound of five common traditional instruments played by 40 artists in seven Dastgahs. In order to address the two mentioned basic tasks, a system is proposed which extracts a sequence of Mel frequency cepstral coefficients (MFCC) feature vectors from input music signal and then converts it to a fixed-length feature vector using i-vector technique. In the classification stage, the extracted i-vector is fed into a support vector machine classifier. The best obtained accuracy on the Nava database for the Dastgah classification and instrument recognition are about 34% and 98% respectively, which indicates the difficulty of the former in comparison with the latter.

In this paper, a Gaussian corrugated profile horn is designed, simulated and compared with a Gaussian profile horn with flat wall, at central frequency of 94 GHz. The proposed horn antenna consists of a corrugated section(mode converter and phase section) and a Gaussian profile section. The Gaussian profile Horn Antenna provides a smooth transition from waveguide to aperture, which improves matching between antenna and free space. This ensures better matching, better radiation pattern, more bandwidth and increases overall system efficiency. The corrugated horn also has low edge diffraction, improved pattern symmetry, and reduced transverse polarization, because these corrugations create same boundary conditions for electric and magnetic fields. The  symmetry pattern on the E and H orthogonal planes, the broad bandwidth as well as low sidelobes of this designed corrugated Gaussian profile horn, make it a good representative for use as a feed in millimeter wave imaging system in W band.

High efficiency video coding standard is the latest standard of international video compression whose first version was released in 2013. In this standard, computational complexity has increased due to the increase in the number of evaluated blocks and prediction modes. Therefore, it is necessary to use methods that reduce the computational complexity and prediction modes. In this paper, in order to reduce the computational complexity, a method has been proposed in which due to the temporal redundancy between frames, the number of evaluated blocks and spatial redundancy, the number of intra-frame prediction modes decreases. First, the block with the highest correlation and similarity to the current block and is around the co-located block within the reference frame is considered as the most correlated block and according to the size of that block, the size of the current block is determined, and evaluating a number of blocks having less chance to be selected is ignored. Then, due to the similarity of pixels in the horizontal or vertical direction, the number of prediction modes decreases. The proposed method is evaluated with 10 different video resolutions having different textures. The simulation results show that the proposed method can reduce the coding time 28.8 % on average, while the BD-rate has a slight increase of 0.9 %.

In 5G wireless communication networks, the non-orthogonal multiple access (NOMA) and the simultaneous wireless information and power transfer (SWIPT) methods can be exploited to improve the spectral efficiency and link reliability, respectively. This paper focuses on a cooperative SWIPT NOMA protocol in MISO downlink case, where a user with strong channel conditions, i.e. user 2 (U2), acts as an energy harvesting (EH) relay to assist a user with poor channel conditions, i.e. user 1 (U1). Furthermore, U2 exploits sun energy harvesting, along with the RF signal energy harvesting through SWIPT, to improve the system performance. A jointly optimizing design is performed on the power splitting (PS) ratio of the relay and beamforming vector of the transmitter to maximize the data rate of U2 subject to satisfying the QoS requirement of U1. We have formulated the optimization problem and evaluated the system performance for different values of harvested sun powers. Our results show that exploiting the solar energy harvesting provides a considerable improvement for the cooperative SWIPT NOMA method in terms of system sum rate and feasible probability of the optimization, especially at low SNR regimes.

The performance of two-way interference-limited amplify-and-forward selective combining relaying systems over independent and non-identically distributed (i.n.i.d.) Nakagami-m fading channels is investigated. Tight lower bounds on the end-to-end outage probability and symbol error rate are derived in closed-form. An asymptotic analysis at high SNRs is done to derive useful expressions for the diversity order and coding gain. Some practical special cases (for example: interference free, infinite power, and Rayleigh fading channels cases) are also studied. Subsequently, an outage minimization problem for three practical scenarios is formulated and solved, analytically. They are power allocation under fixed relay location, relay position optimization under fixed power allocation and joint power allocation and relay position optimization problems. The numerical results confirm the correctness and accuracy of derivations and provide important physical insights into the impact of model parameters on the system performance. For instance, it is demonstrated that the first and third optimization scenarios offer significant performance enhancement over the second optimization scenario.

In this paper, three relay selection protocols for multi-cell networks with multiple decode-and-forward relays targeted by multiple interferers at the relay and the destination are proposed and analyzed. The following three protocols are considered: 1) opportunistic best-worst (OBW) where the destination selects the best signal between the direct and the relayed transmissions, 2) best-worst direct link combiner (BWDC) where the destination combines the direct and the relayed transmissions using maximum ratio combining, and 3) hybrid opportunistic best-worst (HOBW) where the destination only selects the relayed transmission when the direct transmission is in outage. For each of these, new closed-form expressions for the outage probability over independent non-identically distributed (i.n.i.d.) Nakagami-m fading channels are derived. New design insights into the outage probability in the high interference regime are obtained. Compared to BWDC, OBW and HOBW offer a lower complexity receiver. Further, we show that OBW and HOBW achieve a lower outage probability relative to BWDC in the high interference regime.

Wireless sensor networks’ characteristics and limitations have led to severe challenges in their security. In this paper, one of the most common attacks, named as the Byzantine attack, has been studied. A sensor network with a mobile data fusion center is considered, and an optimized data collecting method is proposed which can apply the Neyman-Pearson criterion in a particular case of hard decision making. It means that keeping the false alarm rate constant, the detection probability is maximized. Using statistical data analysis, a closed model is obtained with an acceptable speed of convergence in a various number of sensors and attackers in the network. Then a novel malicious sensor detection scheme is exploited to identify the attack strength using the Hamming distance between every sensor report and the final decision. The simulation results show the superior performance of applying the proposed method in the decision process and also the proposed attack detection method in comparison with the conventional methods.

The use of multi-input multi-output (MIMO) systems in addition to increasing capacity, reducing the destructive effects of multi-path phenomena, reducing interference with other users, will lead to higher data rates. On the other hand, the use of millimeter-wave technology and work in high-frequency bands can prevent issues such as traffic and interference, and can significantly increase the data rates, spectral efficiency and the bandwidth. The millimeter-wave massive MIMO with the lens antenna array can significantly reduce the number of radio-frequency chains. In this paper, two novel algorithms are proposed for channel estimation in millimeter-wave massive MIMO. In this regard, a new algorithm using the compressive sensing based on the convex optimization is presented for channel estimation with high accuracy and low complexity. Then, the second new algorithm based on the greedy methods is provided. One of the benefits of this algorithm is its reduced computational complexity, and its high recovery speed. Finally, both proposed algorithms are compared with other existing algorithms. The simulation results confirm that the proposed algorithms outperform the existing algorithms.

With the rapid development of information, the security of information has gained more and more attention. Information hiding is a useful method to protect secret data through the internet. The High Efficiency Video coding (HEVC) standard is the latest video compression standard ever which can be used to carry hidden data. This paper presents a new method for hiding data in compressed videos using the HEVC standard. In the proposed method, the motion vectors of the inter prediction are used as data carriers. A set of motion vectors of an prediction unit in the HEVC standard is selected to carry a string of hidden data and only a component of that set increases or decreases by one unit, which does not change the peak signal-to-noise ratio (PSNR) significantly, and bit rate does not increase significantly after embedding of hidden data. The results of the experiments demonstrate that the average signal-to-noise ratio decreased by 0.21 decibel and the bit rate increased by an average of 0.76%. Comparison of the proposed method with other recently suggested methods indicates the superiority of the proposed method.

Short packet transmission is one of the topics considered in the new generation of telecommunication systems for low latency connections. In this paper, in order to have a low latency communication, a cooperative system with finite block length (FBL) transmission is investigated. In this relay system, a two-way half-duplex (HD) relay replaces data packets between two nodes. The two-way relay functionality is assumed as amplify-and-forward (AF) with fixed gain and variable gains. Due to the importance of delay, the criterion for evaluating system performance is the effective capacity. Therefore, to optimize the performance of the system, the optimal power allocation is used to maximize the effective capacity. The concavity of this power allocation problem under the assumption of constant total available power between nodes and relay is calculated. Finally, the effect of various parameters including packet length, bit error rate and quality-of-service (QoS) is investigated on overall system performance.

Energy harvesting (EH) (or scavenging) is introduced recently as a prominent technology for alleviating energy or power delimitations of next generation wireless networks, such as industrial systems based on the internet of things (IoT), and also as a bases for the green communication networks, to reduce energy consumption in man-made activities. In this paper, cooperative transmission is considered in full-duplex Gaussian relay channel with energy harvesting source (S) and relay (R) nodes. This paper aims at maximizing total number of transmitted bits from S to destination (D) in a determined time through optimal rate and power allocation to the EH nodes. Also, based on the nodes’ capabilities to transfer some part of their energy (power) to each other, two cases are considered: one-way power transfer (PT) from S to R, and two-way PT between S and R. Optimal resource allocation is investigated to find algorithmic solution. For this purpose, conditions on nodes’ EH profile are considered, where optimal algorithmic solution is obtained in one-way PT case. In the case of two-way PT, optimal solution in general form is presented. Finally, the performance of the proposed algorithms is analyzed numerically and compared with numerical convex optimization tools.

This paper presents a reversible data hiding method in encrypted image that employs correlation of neighboring pixels in the image. In the proposed method, original image may be encrypted by desire encryption algorithm. More significant bits of the pixels in the image are exploited to vacate room for embedding data bits. In the approach, image is divided into separated blocks and most central pixel of each block is considered as reference one. The prediction error between the intensity of other pixels and reference one is calculated and denoted local prediction. This error is analyzed determining a feature of block embedding capacity. Calculated features for all blocks are compressed employing arithmetic coding and embedded in the image along with data bits. At the recipient, at first, compressed features are extracted, then they are uncompressed and used to lossless reconstruction of the original image and extraction of the data bits. Experimental results confirm that the proposed algorithm outperforms state of the art ones.

Orbit Angular Momentum is a novel technology which has been used recently for increasing the channel capacity of communication systems.  In the OAM system, the Uniform Circular Array (UCA) has been widely used due to its capability in easy implement of the OAM modes orthogonality.In most of the researches, it is assumed that the transmitter and receiver antenna arrays are aligned with each other in the channel without any reflection. The effect of channel reflection is investigated only on single antenna and Uniform Linear Array (ULA) cases, which according to many difference between the UCA  and  ULA based OAM systems, the obtained results cannot been used for UCA based OAM system. In this paper, the channel modeling of the UCA based OAM system in presence of the reflection in done. For this purpose, the reflection effect is modeled by a virtual transmitter UCA. Also, the channel matrix in extreme cases is approximated, and the channel capacity using the obtained model is computed. We show that, in the presence of channel reflection, the orthogonality of OAM modes are corrupted, which leads to decreasing the channel capacity. Consequently, a proper pre-coding is needed to prevent capacity loss. We evaluated the obtained results by the simulations.