per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2021-10 18 2 3 28 article Maryam Ebrahimi m.ebrahimi@itrc.ac.ir 1 Mohammad Hesam Tadayon tadayon@itrc.ac.ir 2 Mohammad Sayad Haghighi sayad@ut.ac.ir 3 Iran Telecommunication Research Center, Iran, Sama Technical and Vocational Training College, Islamic Azad University Iran Telecommunication Research Cente School of Electrical and Computer Engineering, University of Tehran In the complex Internet of Things (IoT) paradigm that things interact with each other as well as with human beings, one approach is to implement trust management systems in order to provide security for smart network applications. Trust, in general, overlaps with concepts such as privacy, security, and reliability. However, the high number of objects in IoT, along with its dynamic nature and existence of malicious entities, make IoT trust management quite challenging. These attributes rule out the possibility of using traditional best practices for IoT networks. Trust management algorithms have been implemented for a variety of applications in IoT environments. These algorithms are usually utilized to enhance the quality of received services in the presence of malicious entities. Such algorithms and methods have been proposed to secure IoT networks in different contexts, including traffic routing, smart cities, vehicular ad-hoc networks, healthcare ecosystems, and object authentication. In this paper, first, different state of the art trust computation methods are numerically evaluated to estimate trust in a common testbed. Finding the best approach to assign a precise value to the trust level of an object is a crucial matter. Therefore, the principal parameters that make trust computation methods different are extracted and then, the existing trust calculation approaches built upon them are categorized. Type of relationship, direct trust, indirect trust, combination of trust values, trust updating process, data storage, and social relationships are considered as the parameters to analyze trust computation models with. Type of relationship between trustor and trustee can be different. Either of them can be object or human. Moreover, trust is usually a combination of direct experiences and recommenders’ feedback. There are different update methods too. Trust estimation can be updated after each transaction, a definite time interval, or both of them. Depending on the storage and accessibility of data, algorithms can be built to be centralized, decentralized or semi-centralized. Moreover, social parameters can be involved in trust assessment, which is the subject of trust management in Social IoT. After analyzing each of these parameters’ effect on trust assessment, in the next part of the article, trust-related attacks are studied. Every method that can make trust management models resistant to attacks is explained.  We introduce relevant attacks and their countermeasures in direct, indirect, and hybrid trust calculation algorithms. More importantly, we study the methods of trust model evaluation and the effect of limited resources on the performance of trust calculation algorithms. In short, we conduct a comparative survey in which trust-related IoT works are studied from four perspectives: (1) Trust calculation principles, (2) Attack resistance, (3) The effect of resource limitation on model performance, and (4) Trust management evaluation framework. Through this, we find the advantages and disadvantages of existing algorithms and make a measure for the evaluation of IoT trust management systems. We provide comparative tables to show the differences between IoT trust models. A major contribution of this paper is establishing quantitative metrics to assess trust estimation models and reveal their strengths and weaknesses under different conditions. http://jsdp.rcisp.ac.ir/article-1-1123-en.pdf Trust Internet of Things Social Internet of Things Trust Evaluation per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2021-10 18 2 29 44 article Hamid Reza Ghaffari hghaffaripaper@ferdowsiau.ac.ir 1 Atena Jalali Mojahed st.ajalalia@ferdowsiau.ac.ir 2 Department of Computer Engineering,Department of Computer Engineering, University of Ferdows, Ferdows, Iran Department of Computer Engineering,Department of Computer Engineering, University of Ferdows, Ferdows, Iran Classification is a machine learning method used to predict a particular sample’s label with the least error. The present study was conducted using label prediction ability with the help of a classifier to create a new feature. Today, there are several feature-extraction methods like principal component analysis (PCA) and independent component analysis (ICA) that are widely used in different fields; however, they all suffer from the high cost of transferring to another space. The purpose of the proposed method was to create a higher distinction between various classes using the new feature in a way that, make the data in the classes closer to each other. As a result, for increasing the efficiency of classifiers, more differentiation is created between the data of various classes. Firstly, the suggested labels for the primary data set were determined using one or more classifiers and added to the primary data set as a new feature. The model was created using a new data set. The new feature for training and testing data sets was provided separately. The tests were performed on 20 standard data sets and the results of the proposed method were compared with those of the two methods described in the related studies. The outputs indicated that the proposed method has significantly improved the classification accuracy. In the second part of the tests, the resolution of the new feature was examined according to two criteria, namely Information Gain and Gini Index, for examining the effectiveness of the proposed method. The results showed that the feature obtained in the proposed method has higher Information Gain and lower Gini Index in most cases, as it has less irregularity. To prevent the increase in data dimensions, the feature with the least Information Gain was replaced with the feature extracted with the most Information Gain. The results of this step showed an increase in efficiency as well. http://jsdp.rcisp.ac.ir/article-1-986-en.pdf Feature extraction classification information gain Gini index per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2021-10 18 2 45 56 article Ehsan Imani far ehsanimanif@gmail.com 1 Amir Akhavan aakhavan@cc.iut.ac.ir 2 ali asghar abniki aliasghar_abniki@yahoo.com 3 Department of Biomedical Engineering-Bioelectric, Amirkabir University of Technology Department of Electrical and Computer Engineering, Isfahan University of Technology Center of Excellence in Hydrodynamics and Dynamic of Marine Vehicles, Department of Mechanical Engineering, Sharif University of Technology Direction of Arrival (DOA) estimation of sound sources using phased array-based methods has a lot of importance in various fields, including sonar, robot vision, and mechanical defect detection. Adaptive beamforming methods, such as the MVDR (Minimum Variance Distortionless Response) algorithm, have high resolution compared to non-adaptive methods (Delay and Sum algorithm); but this advantage is achieved in return for the computational complexity of these algorithms. This makes it hard to use these algorithms in applications that require real-time sound source DOA estimation. On the other hand, an important feature of the adaptive beamforming methods including MVDR is the high potential of these algorithms for parallelization. The purpose of this paper is the parallel implementation of the MVDR algorithm by employing Graphical Processor Unit (GPU) instead of Central Processor Unit (CPU) to increase the execution speed and achieve the real-time mode. For this purpose, the CUDA (Compute Unified Device Architecture) programming model has been used to implement the algorithm on the GPU. CUDA is a parallel computing platform and application programming interface (API) model created by Nvidia. It allows software developers to use a CUDA-enabled GPU for parallel processing. In order to investigate the performance of parallel implementation of the MVDR algorithm, two different GPUs, as well as CPUs, have been used. The performance validity of various implementations in this paper was confirmed by real sonar data as well as simulation data. The results show that using an array of 64 sensors, it is possible to estimate the DOA of underwater sound sources in real-time mode and with high resolution using the MVDR algorithm. http://jsdp.rcisp.ac.ir/article-1-990-en.pdf DOA estimation of sound sources MVDR algorithm Parallel processing GPU CUDA per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2021-10 18 2 57 74 article mina Abbaspour orangi Mina.Abbaspour@aut.ac.ir 1 Alireza Hashemi golpayegani Sa.Hashemi@aut.ac.ir 2 Department of Computer Engineering and Information Technology, Amirkabir University of Technology Department of Computer Engineering and Information Technology, Amirkabir University of Technology Trust is one of the most important cornerstones in social networks' discussions. mostly the way that users of these networks trust each other are considered identical, while these users can have different approaches and considerations in trusting others. Meanwhile, users can impress each other and change their trusting patterns in other users. As a result, the mechanism and manner of impressing opinion trust behavior and conditions of behavioral modes changing have a place of importance to be considered. The question is that, how we can consider different behavior of users and their impression in trusting others? In the first step, the main purpose of this paper is to spotlight social networks' different user behavior in trusting other users. For this purpose, the three most important behavioral modes in  users trust are considered. In each of these modes, behavioral and functional characteristics of users are the basis of calculating trust, which is based on mental beliefs of them. These modes are named as optimistic, moderate, and pessimistic trusting modes. In optimistic mode, we suppose that users think positively and consider low level of activities and signs in trusting others. Here, negative interactions have little impact on users mind. In moderate mode, we suppose that users are not as optimistic as mode A and consider all the interactions and signs when they want to trust others. Here, any negative action can destroy the trust of users and has a greater impact on users. Finally, in pessimistic mode, we suppose that users are pessimistic and hardly trust someone. In this mode, the interactions that happened more recently have more value than those happened in the past. In the next step, the purpose and innovation of this paper is the way that the trust behavior of users spreads. Three different scenarios are considered for the impressing and spreading of nodes behavior, purposely. In each scenario, different states for users and different purposes for diffusion are defined.   Next, it is followed by maximizing of impression and finding more impressive agents in diffusing trust behavior through social networks. For this purpose, it's focused on the structure of users social networks, and the most impressive ones are determined through different diffusion scenarios. The findings of this article appear a significant discrepancy in the amount of trust in each of the different behavioral modes, which is more acceptable in the real world. Analyzing test results leads us to the fact that in the presented model, choosing the start node from each community with 48.14 percent in behavior improvement and diffusion speed and the nodes with the highest degree with 37.03 percent in behavior changing has much more reasonable results than usual models. http://jsdp.rcisp.ac.ir/article-1-967-en.pdf Trusting behavior Trust diffusion Influential nodes Social networks Social network analysis per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2021-10 18 2 75 96 article Maryam Yousefpour Maryam.yousefpour@yahoo.com 1 Behrang Barekatain behrang_barekatain@yahoo.com 2 University University One of the most widely researched areas in wireless sensor networks is the issue of coverage. Coverage is directly related to the degree of quality, how and how long the sensors are able to detect predetermined targets in areas. Standard coverage for calculating network Quality of Service (QoS) greatly affects network performance. The most important factor in maximum coverage in wireless sensor networks is to consider the energy consumption constraints. Sensor nodes typically use a battery to supply power, which in most cases is not rechargeable; therefore, it is very important to carry out operations to reduce energy consumption and increase the life of the sensor node. Full network coverage is such that all areas of interest must be monitored and covered by diffused sensors. Now, if a certain area is not covered by any sensor, it is known as a cover hole and the real events in the cover points will not be recognizable and traceable. The aim of this study is to identify cover holes to reduce energy consumption in wireless sensor networks; which is applied by presenting a new model of credit management system. In this system, the sensor radius of the nodes is adjusted for maximum coverage in the network. In the simulation presented, decisions are made by gathering information at different time intervals and interactions among nodes; According to this decision, the evaluation is applied for the level of satisfaction of the nodes and based on the level of satisfaction of the nodes, rewards and fines are considered for the nodes and finally, using the support vector machine tool, cover holes in the sensor network are identified. Thus, the effect of identifying holes in reducing energy consumption, performance and quality of services provided in the network is expressed. Evaluations show that the longevity of the network and the size of the sensor radius of the nodes have improved 12.42% and 20.4%, respectively in terms of the number of sensor nodes in the environment with respect to the moving radius of the nodes, mobile nodes and cellular automata compared to the environment considering the constant radius of the nodes, mobile nodes and without automata. http://jsdp.rcisp.ac.ir/article-1-1014-en.pdf Coverage Cavities Wireless Network Sensing Radius Trust Management Energy Consumption Support Vector Machine per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2021-10 18 2 97 114 article musa mojarad mosa.mojarad@gmail.com 1 hamid parvin parvin@iust.ac.ir 2 samad nejatiyan samad.nej.2007@gmail.com 3 Karam allah Bagheri Fard k.bagheri@iauyasooj.ac.ir 4 Department of Computer Engineering, Firoozabad Branch, Islamic Azad University Department of Computer Engineering, Nourabad Mamasani Branch, Islamic Azad University Department of Electrical Engineering, Yasooj Branch, Islamic Azad University Department of Computer Engineering, Yasooj Branch, Islamic Azad University Background: There are many theories about the causes of hereditary diseases, but physician believe that both the genetic and environmental factors simultaneously play an important role in the development and progression of these diseases, although the extent to which this effect is not yet clear. In order to detect effective genes in the development of diseases, it is necessary to achieve the relationship between cells/tissues. Objective: In fact, inter-cell or inter-tissue communications indicate the hereditary relationships between patients. Detecting these communications help to identify common parts of the body that are influenced by various diseases. The interaction between different cells/tissues can be demonstrated by expressing the gene between them. By sampling chromosomes, useful information is obtained about the type of disease and how it is transmitted. By examining this information, you can identify disorders that have led to highly altered changes. In previous research, various clustering methods have been used to discover the links between diseases based on gene expression data. However, ensembl clustering approaches have not yet been used for this purpose. Method: In this paper, the recognition of intercellular and inter-tissue interactions in various diseases have been done according to the characteristics of the topological structure of the graph and an improved ensembl clustering method. The proposed clustering algorithm uses an agreed similarity function to measure the similarity between objects. The proposed method has two stages; in the first step, several clustering models are combined to identify the initial relationships between cells or tissues in order to produce better results than individual algorithms. In the second stage, the similarity between cells or tissues in each cluster is calculated by using a similarity criterion based on the topological structure of the graph. Eventually, the maximum similarity between cells or tissues in each cluster is used to discover the relationship between diseases. In addition, an algorithm for improving the uncertainty of objects is evaluated by allocating them to other clusters in order to enhance the quality of the final clusters. Results: To evaluate the performance of the proposed method, several UCI datasets and the FANTOM5 dataset have been used. The results of the proposed method on the phantom data set 5 report a silhouette of 0.901 in 18 clusters for cells and 0.762 in 13 clusters for tissues. Conclusion: The conducted evaluations have confirmed the power of the proposed clustering algorithm in terms of accuracy. Clustering of cells or tissues has increased the accuracy and concentration of the topological similarity criterion of the graph in the range of similarity of cells or tissues. http://jsdp.rcisp.ac.ir/article-1-980-en.pdf Intercellular communication Improved clustering Graph topological structure FANTOM5 dataset per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2021-10 18 2 115 134 article Maedeh Daryanavard shahbahrami@guilan.ac.ir 1 Asadollah Shahbahrami shahbahrami@guilan.ac.ir 2 Department of Computer Engineering, Engineering Faculty, Islamic Azad University, Lahijan Department of Computer Engineering, Engineering Faculty, University of Guilan Objective Image Quality Assessment (IQA) algorithms are divided into three categories according to the availability of the reference image and the amount of information available in the assessment, namely, algorithms with Full Reference (FR), Reduced Reference (RR), and No Reference (NR). If the original high-quality image exists in the evaluation algorithm at the same time to compare with the test image, and both images are identical in content, the evaluation is called FR assessment. If only a few features extracted from the high-quality image are used to compare with the test image, it is called RR assessment. In NR assessment algorithms, there is no feature or reference image to compare with the test image. Algorithms in NR are divided into two subcategories, Distortion-Specific (DS) and Non-Distortion-Specific (NDS). In first one, algorithms predict the quality of an image by knowing the type of distortion which is effective when distortion information or type is available. However, information about the type of distortion is not available in most applications which limits the use of these algorithms. The NDS algorithms can be applied to different types of distortion and are designed to be all-purpose. The NDS algorithms are divided into two subcategories namely, Opinion Aware (OA) and Opinion Unaware (OU). In the OA model, the images are evaluated and scored by the human factor, and each image with its corresponding human score is mapped by the learning system, while the OU model does not have the score of a human observer and is evaluated completely blind. The proposed algorithm is this paper is for NR model and NDS and opinion unaware. Image quality is well correlated with features of local structure, contrast, and color. By modeling, these features with distributions such as Gaussian or Gaussian families can be used to detect image degradation. The proposed method consists of two stages of training and testing. Five NSS features that are actually extracted from the MSCN coefficient. The q-Gaussian distribution model is used for image distribution. The q-Gaussian distribution is one of the options that create flexible decision boundaries with different Gaussian shapes that are more generalizable in anomalies than other distributions. The learning phase is performed only once to extract the features of images and to be considered as a model in the system to compare with the images that are to be entered as test images. To evaluate the performance, the proposed method is compared with IL-NIQE which is similar to the proposed method in terms of behavioral mechanism and it uses natural scene statistics. Performance metrics such as PLCC, KROCC, RMSE, SROCC and some datasets, LIVE, CSIQ, and TID2013 have been used for evaluation. The proposed method performs better than the compared technique. The proposed method can show better performance due to the adjustable parameter of q. http://jsdp.rcisp.ac.ir/article-1-1050-en.pdf Image quality assessment No-reference assessment Non-distortion-specific assessment q-Gaussian distribution per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2021-10 18 2 135 146 article Fatemeh Mirjalili f.mirjalili93@ms.tabrizu.ac.ir 1 Jafar Razmara razmaraj@gmail.com 2 University of Tabriz Department of Computer Science, Faculty of Mathematics, Statistics, and Computer Science, University of Tabriz In recent years, the speed and complexity of computer networks have grown significantly. At the same time, network-based anomalies and attacks have increased. Nowadays, intrusion detection and prevention is considered as a main strategy in satisfying the security of computer systems and communication networks, and the detection of these attacks with high accuracy and the least error is very important, especially in the field of network management. Today, virtualization technology is widely developing in order to set up multiple virtual systems on a physical system. Computational clouds are the most hallmark of this technology. Intrusion detection systems play a key role in protecting cloud resources on virtual machines. An intrusion detection system has the task of monitoring events within a computer system and the communication networks, and detects unauthorized and abnormal behaviors to deal with them. The proposed systems for intrusion detection mainly use data mining, machine learning and statistical analysis of data. Therefore, it is natural that in some cases they lead to the production of false alarms. Consequently, it is essential to improve the accuracy and high detection capability of these systems. Regarding the increasing speed and complexity of these machines, it is necessary to increase the ability and accuracy of intrusion detection systems for identifying different types of attacks at a right time. In this regard, the use of behavior-based approaches has attracted more attention due to their high scalability in the large networks. The methods for intrusion detection that utilize network traffic graph clustering do not have the accuracy and appropriateness with the speed of data transfer in the current computer networks. Thus, the solutions can be improved by choosing an appropriate strategy for clustering. In this paper, a new behavior-based method for detecting intrusion in computer networks is presented. To this end, the network data was modeled through the flow of data as a traffic dispersion graph and then clustered using an improved Markov-based algorithm. Then, by analyzing a set of statistical criteria, the produced clusters, a penetration detection model was constructed. A set of modified statistical criteria was defined and utilized for analyzing the constructed clusters. The proposed model was examined and evaluated on the DARPA 99 dataset. In addition, the results of the proposed method were compared with seven other methods which work based on machine learning techniques. The results show that in the proposed method, the error detection rate is significantly reduced and the accuracy rate of the method is increased compared to seven other intrusion detection approaches. The reason for this performance improvement can be attributed to the good performance of Markov's improved clustering algorithm, which has produced more accurate results on flow-based data. Also, defining and applying appropriate criteria to determine the threshold limits is effective in obtaining accurate results. In addition, the results demonstrate that the proposed model has better capabilities than the methods which are not use graph clustering and can detect attacks with high accuracy. http://jsdp.rcisp.ac.ir/article-1-984-en.pdf behavior-based intrusion detection traffic dispersion graph data flow graph clustering optimized Markov clustering per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2021-10 18 2 147 162 article Zahra Hossein-Nejad hoseinnejad.zahra@yahoo.com 1 Mehdi Nasri nasri_me@iaukhsh.ac.ir 2 Department of Electrical Engineering, Sirjan Branch, Islamic Azad University Department of Electrical Engineering, Khomeinishahr Branch, Islamic Azad University Image mosaicing refers to stitching two or more images which have overlapping regions to a larger and more comprehensive image. The image mosaicing process is widely used in scene stabilization, change detection, video compression, and image compression. Image mosaicing methods can be divided into two categories, direct methods and feature-based methods, which feature-based methods are more accurate. Scale Invariant Feature transform (SIFT (is one of the most common feature-based methods in the image mosaicing. However, one of the big defects of SIFT algorithm is the large number of duplicate key points and being time-consuming due to the high dimensions of classical SIFT descriptor. In this paper, to solve these problems, a new four-step approach for image mosaicing is proposed. At first, Redundant Keypoint Elimination-SIFT (RKEM-SIFT) algorithm which has been proposed in [1] is used to identify keypoints of reference and sensed images and to improve the mosaicing process. In the second stage, for each keypoint of the image, 64-D SIFT descriptor is computed. In this descriptor, unlike the 128-D SIFT descriptor, a smaller window is used which improves the accuracy of matching and reduces the running time. In the third stage, the proposed improved RANdom SAmple Consensus (RANSAC) algorithm is used to determine the adaptive threshold in the RANSAC algorithm to remove the mismatches and to improve the image mosaicing. Determining the appropriate threshold value in RANSAC is so important, because if an appropriate value is not chosen for this algorithm, the mismatches are not removed, and eventually there will be a serious impact on the outcome of the image mosaicing process. In this method, the threshold value is based on the median value of distances between matching points and their transformed model. Image blending in the mosaicing process is the final step which blends the pixels intensity in the overlapped region to avoid seams. The suggested method of blending is to combine the images based on the average of the data in the overlapped region of two images. The proposed blending method reduces artifacts in the image for better performance of the mosaicing process. Another advantage of this proposed method is the possibility to combine more than two images that are suitable for creating panoramic images. The simulation results of the proposed image mosaicing technique, which includes the RKEM-SIFT algorithm as feature detector, 64-D SIFT descriptor, proposed adaptive RANSAC algorithm, and proposed image blending algorithm. The proposed method is implemented on standard image databases, created image databases, and has been compared with SURF- fast bidirectional matching, SURF-LM and SIFT-RANSAC methods. The results of the experiments show the superiority of the proposed method according to the criteria of mean square error and accuracy, which compared to the best compared method (SURF-fast bidirectional matching) reduces 6.7% maximum error, 30.09% root mean square error and 37.68% caused the median error. http://jsdp.rcisp.ac.ir/article-1-1008-en.pdf Image mosaicing Image registration SIFT per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2021-10 18 2 163 176 article Mahsa Vaghefi mahsavaghefi@gmail.com 1 Fatemeh Jamshidi Jamshidi@fasau.ac.ir 2 Department of Electrical Engineering, Shiraz Branch, Islamic Azad University Department of Electrical Engineering, Fasa University, Fasa, Fars province Any heart activity disorder may lead an irregularity in is rhythm, or cardiac arrhythmia. An ECG signal is one of the major tools for classifying different types of cardiac arrhythmias. ECG signals usually contain various noises. To have a better signal processing, it is essential to remove noises in a way that a signal structure never becomes subject to distortion. After the step of noise removal, selection of an appropriate method is of paramount importance for feature extraction. Optimal features can be selected to improve efficiency and reduce calculations. This article used the ensemble empirical mode decomposition (EEMD) in which any intrinsic mode function (IMF) contains only a single frequency component for noise removal. The noise removal operation with the least distortion is possible using an appropriate windowing on a QRS complex containing sum of the first three IMFs. Later, the remaining noises can be removed using discrete wavelet transform (DWT). The results of using the EEMD-DWT combined method were compared with EMD and DWT combination. After the noise removal step, feature extraction was performed through a wavelet packet decomposition. It is capable of signal decomposition at all frequencies. Multiple objective binary particle swarm optimization (MOBPSO) method was used to select optimal features and the effect of this method on the results was examined. Finally, the back propagation neural network (BPNN) and a support vector machine based on particle swarm optimization were used for classification. This article used 17 signals received from the MIT-BIH database. The acquired data belong to 6 different types of classes. After pre-processing, feature extraction, feature selection, and classification on the input data, it is observed that the proposed technique of EEMD-DWT is an appropriate method for noise removal and MOBPSO is a suitable method for the selection of best features. The BPNN classifier managed to classify cardiac arrhythmias with a higher accuracy and the values for accuracy, sensitivity, specificity, and positive predictive value were 99.12%, 97.08%, 99.38%, and 97.12%, respectively. http://jsdp.rcisp.ac.ir/article-1-972-en.pdf Back Propagation Neural Network Discrete Wavelet Transform Electrocardiogram Ensemble Empirical Mode Decomposition Multiple Objective Binary Particle Swarm Optimization