per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 1 18 article payam bahrani p.bahrani@gmail.com 1 Behrouz Minaei Bidgoli b_minaei-at@ac.ir 2 Hamid Parvin parvin@iust.ac.ir 3 Mitra Mirzarezaee mitra_mirzaee@gmail.com 4 Ahmad Keshavarz ahmad_keshavarz@gmail.com 5 1Department of Computer Engineering, Science and Research branch, Islamic Azad University, Tehran 2Department of Computer Engineering, Iran University of Science and Technology, Tehran 3Department of Computer Engineering, Nourabad Mamasani Branch, Islamic Azad University, Nourabad Mamasani, IR 5Department of Electrical Engineering, Persian Gulf University, Bushehr 5Department of Electrical Engineering, Persian Gulf University, Bushehr Recommender systems that predict user ratings for a set of items are known as subset of information filtration systems. They help users find their favorite items from thousands of available items. One of the most important and challenging problems that recommendation systems suffer from is the problem of dispersion. This means that due to the scatter of data in the system, they are not able to find popular items with the desired reliability and accuracy. This is especially true when there are a large number of items and users in the system and the filled ratings are low. Another challenging problem that these systems suffer from is their scalability. One of the major problems with these systems is the cold start. This problem occurs due to the small number of items rated by the user, i.e. the scatter of users. This problem is divided into two categories: new user and new item. The main focus of this article is on the problem of the new user type. This problem occurs when a new user has just logged in and has not rated any item yet, or when the user has already logged in but has been less active in rating. The goal is to address these three challenges. In this study, an ontology-based hybrid recommender system is introduced in which ontology is used in the content-based filtering section, while the ontology structure is improved by the collaborative filtering section. In this paper, a new hybrid approach based on combining demographic similarity and cosine similarity between users is presented in order to solve the cold start problem of the new user type. Also, a new approach based on combining ontological similarity and cosine similarity between items is proposed to solve the cold start problem of the new item type. The main idea of the proposed method is to extend users’/items’ profiles based on different mechanisms to create higher-performance profiles for users/items. The proposed method is evaluated in a real data set, and experiments show that the proposed method performs better than the advanced recommender system methods, especially in the case of cold start. http://jsdp.rcisp.ac.ir/article-1-1199-en.pdf Recommender System Ontology Profile Expansion Hybrid Recommender System per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 19 38 article Hossein Sadr Sadr@qiau.ac.ir 1 Mir mohsen Pedram Pedram@khu.ac.ir 2 Mohammad Teshnehlab Teshnehlab@eetd.kntu.ac.ir 3 Department of Computer Engineering Rasht Branch, Islamic Azad University Rasht, Iran Department of Electrical and Computer Engineering Faculty of Engineering, Kharazmi University Tehran, Iran Industrial Control Center of Excellence Faculty of Electrical and Computer Engineering, K. N. Toosi University Tehran, Iran People's opinions about a specific concept are considered as one of the most important textual data that are available on the web. However, finding and monitoring web pages containing these comments and extracting valuable information from them is very difficult. In this regard, developing automatic sentiment analysis systems that can extract opinions and express their intellectual process has attracted considerable attention in recent years. Sentiment analysis is considered as one of the most active research areas in the field of natural language processing which tries to classify a piece of text containing opinions based on its polarity and determine whether an expressed opinion about a specific topic, event or product is positive or negative. Since about a decade ago, many studies have been carried out to investigate the effects of traditional classification models, such as Support Vector Machine (SVM), Naïve Bayes, Logistic Regression, etc. in the task of sentiment analysis. Although machine learning models have achieved great success in this filed, they are still confronted with some limitations, notably manual feature engineering requirements. In other words, the classification performance of machine learning models is highly dependent on the extracted features and they play an important role in obtaining higher classification accuracy. To deal with these problems, deep learning models have been extensively employed as an alternative to traditional machine learning models and have achieved impressive results. It is worth mentioning that despite the remarkable performance of these methods, they are still confronted with some limitations and they are on their first steps of progress. Therefore, the goal of this paper is to propose a combinational deep learning model that can overcome their problems as well as utilizing their benefits. In this regard, an efficient method based on combination of convolutional and recursive neural networks is proposed in this paper that employs a generalized recursive neural network, where an intermediate feature is obtained by combining children's nodes, as an alternative of pooling layer in attention-based convolutional neural network with the aim of capturing long term dependencies and decreasing the loss of local information. Based on empirical results, the proposed method with the accuracy of 53.92% and 92.89% respectively on SST1 and SST2 datasets not only outperforms other existing models but also can be trained much faster. http://jsdp.rcisp.ac.ir/article-1-1060-en.pdf Sentiment analysis Deep Leaning Convolutional neural network Recursive neural network Attention mechanism per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 39 42 article Hadiseh Pourali h.poorali@nit.ac.ir 1 Hesam Omranpour h.omranpour@nit.ac.ir 2 Babol Noshirvani University of Technology Babol Noshirvani University of Technology In this paper, a method is presented for predicting time series. Time series prediction is a process which predicted future system values based on information obtained from past and present data points. Time series prediction models are widely used in various fields of engineering, economics, etc. The main purpose of using different models for time series prediction is to make the forecast with the greatest accuracy. The model presented in this paper is based on the combination of kernels and support vector regression. Support vector regression is highly capable of solving function estimation problems by using its kernels, but kernels’ parameters need to be adjusted. First we have preprocessing phase which includes normalizing data and separating data for testing and training. In proposed model, ten different kernels were used. Five kernels were selected as the best kernels by trial and error and these kernels are applied to data. There probably is only a few of the kernels that are useful for the problem, and we are not aware of which kernels are useful for our problem so kernel outputs aggregate by applying a coefficient. This combination creates a new secondary space. The output is given to support vector regression to construct a model that predicts values exactly ɛ accurate, which means the predicted values do not deviate more than ɛ from the original data. This model predicts values by using a leave one out model. Each kernel has parameters that need to be set to optimum values in order to get the best results. Hence in the proposed model, the kernel parameters and their weights are learned by the Gray Wolf Optimizer. This optimizer has been able to provide appropriate answers to many problems, especially challenging problems and has a superior ability to solve the high-dimension problems. By running program in consecutive iterations and examining the different values of the parameters, the optimizer learns the best of them which prediction error has been reduced, and finally returns their best value. The proposed model is implemented on five standard time series and compared to other method, test based on the RMSE criterion for DJ time series, improved by 1.58 point, Radio time series, improved by 0.178 point, and Sunspot time series, improved by 1.709 point. Finally, we analyzed the results, Statistical evaluation by Wilcoxon Signed-Rank Test where the p value is very low compared to the proposed method and CNN-FCM, AR_ model per scale, Multiresolution AR model and ANN methods, slightly lower for Wavelet-HFCM and ANFIS methods and slightly lower than one for SAE-FCM method and at the end provide a relation to find the window size in the model by obtaining the average of peak differences, valley differences, and consecutive peak, and valley differences for the actual values of the training data in exchange for their sequence number in time series. http://jsdp.rcisp.ac.ir/article-1-1162-en.pdf Time series prediction Support vector regression Ensemble kernel model Optimization per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 43 58 article Abbas Dehghani Dehghani.abas@gmail.com 1 Keyvan RahimiZadeh RahimiZadeh@gmail.com 2 Yasouj University Yasouj University Network-on-Chip (NoC) has emerged as leading interconnection backbone to integrate numerous blocks in a single chip. Although it offers a high-performance communication infrastructure by using integrated switch-based networks, the possible performance improvement of a conventional NoC is restricted by multi-hop communications due to high transmission latency and power consumption incurred by the data transmission between two distant cores.  In order to mitigate this problem, wireless NoC (WNoC) architecture has proposed as an alternative solution to design flexible, low-power, and high bandwidth communication infrastructures for the future multicore platforms. It is necessary to mention that wire-based interconnections are still highly effective for short distances communications. Therefore, hybrid WNoC architectures are emerged as scalable communication structure to alleviate the deficits of traditional NOC architecture for the modern multicore systems. The hybrid WNoC architecture provides energy efficient, high data rate and flexible communications for NoC architectures. In these architectures, each wireless router is shared by a set of processing cores. However, sharing links between cores increases congestion in the network that limits the performance and scalability of NoCs and affects the system to work at less than its peak gain. Moreover, the congestion can heightens network inefficiency when the network is scaled to more nodes. In this paper, we propose a novel congestion-aware mesh-based WNoC architecture to address these issues. We consider optimization of the system cost and performance, simultaneously. For congestion control, it is recommended to include a multi-path routing. This means that several routes are calculated and recorded for each destination and finally the traffic load is distributed. Paths are selected based on their scores, which are obtained dynamically. When a path is used to transmit packets, the score of that path is reduced so that fewer packets are sent from that path and more scored paths are used. This approach aims to the distribution of traffic loads on the paths. The performance of the proposed architecture has been evaluated and compared with notable WNoC architectures through comprehensive simulations. The experimental results demonstrated the effectiveness of the proposed design under both synthetic and realistic traffic patterns in terms of network throughput, latency, and energy consumption. http://jsdp.rcisp.ac.ir/article-1-1077-en.pdf Network on Chip Wireless communications Multicore System-on-Chip Congestion per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 59 74 article Mohsen Rezvani mrezvani@shahroodut.ac.ir 1 Mansoor Fateh mansoor_fateh@shahroodut.ac.ir 2 Shahrood University of Technology Shahrood University of Technology The expansion of the use of information exchange space and public access to communication networks such as the Internet has led to the growing dependence of social institutions on the use of these networks. However, maintaining the security of information exchanged on networks is one of the most important challenges for users of these networks. One way to protect this data is to use private networks. But building these networks is not cost-effective in terms of time and cost. In contrast, the use of encryption techniques, access control mechanisms and data concealment are among the effective solutions for security in the information exchange space. Existing methods for hiding information can be divided into three categories: cryptography, watermarking and steganography. In cryptography, a simple text is converted into encrypted text, which, of course, requires a decryption operation as well as an encryption key. In general, cryptographic techniques suffer from two major problems. The first problem is the ban on the transmission of encrypted data in dictatorial regimes, and the second problem is that cryptographers pay attention to encrypted data and stop any secret communication. The second category of information hiding methods is watermarking. Watermarking techniques are commonly used to protect the copyright of a digital content and to deal with issues such as fraud, fraud and copyright infringement in the data transfer space. In steganography methods, the transfer of information takes place in a cover through public communication channels, and only the sender and receiver are aware of the existence of a secret message. Two aspects of steganography must be observed. The first aspect is that the cover and secret content look the same in the face of statistical attacks. The second aspect is that the process of hiding the secret message in the cover is such that there is no difference between the cover and the secret in terms of the human perceptual system. In fact, the accuracy of the transmission media is maintained. Steganography methods use image, video, protocol, audio, and text platforms to hide information. Steganography in the text is difficult due to very little local variation. Humans are very sensitive to textual changes. Hence it is difficult to spell in the text. However, due to the high use of text in digital media, the insensitivity of text to compression, the need for less memory to store and communicate more easily and faster, many methods for steganography have been introduced in it. In addition, text is still one of the major forms of communication available to the general public around the world. In this paper, we propose a new email steganography scheme using a dictionary-based compression. In the proposed scheme, a number of email addresses containing a hidden message will be generated using the submitted text. The submitted text is sent to the generated and recipient addresses at the same time. This does not reveal the identity of the recipient of the message, and only the recipient can extract secret message using other email addresses. In the proposed method, two steganography keys are used. Using these two keys increases the security level of the proposed method. Also, the capacity of the proposed method is unlimited, which of course is a great advantage in a steganography method. This unlimited capacity provides high security for the proposed method. Another advantage is that the proposed method is not limited to the type of the cover-text. Initially, the secret message is converted to a bit string by a dictionary. Then the operation of embedding the secret message in the recipient's addresses is done by the steganography keys. The efficiency of steganography algorithms depends on various factors such as lack of detection by the human eye, lack of detection by statistical methods, and capacity. The proposed method does not change the cover-text. Hence, this method is not detectable by humans or statistical methods. The capacity of the proposed method in this research is based on built-in email addresses. As the text of the message increases, the number of emails created increases too. Of course, this increase in the address of the emails created can lead to suspicion of the emails sent. Therefore, the parameter of the number of emails created is also important in the evaluation. In this paper, the efficiency of the proposed method is evaluated based on the two parameters and compared with existing methods. The results of this evaluation show that the proposed method, in addition to providing unlimited capacity in steganography, produces fewer email addresses generated as well as fewer message bits after compression. http://jsdp.rcisp.ac.ir/article-1-1047-en.pdf Email Steganography Dictionary Capacity Security per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 75 86 article fershteh javadzadeh fereshteh.jvz@gmail.com 1 mehdi yaghoubi mehdi.yaghoubi@gmail.com 2 soheila karbasi s.karbasi@gu.ac.ir 3 In recent years, business process management (BPM) has been highly regarded as an improvement in the efficiency and effectiveness of organizations. Extracting and analyzing information on business processes is an important part of this structure. But these processes are not sustainable over time and may change for a variety of reasons, such as the environment, human resources, capital market changes, seasonal, and climate changes. These changes in business processes are referred to as concept drift in event logs. The discovery of concept drifts is one of the challenges in business process management. These drifts may occur suddenly, gradually, periodically, or incrementally. This paper proposes an algorithm for identifying sudden concept drifts in event logs that are created by BPM. Each execution of the process instance follows a specific path in the process model called a trace, all traces that follow the same path in process model are called a variant. The proposed algorithm is based on the distribution of trace variants in the execution of processes. In this method, by moving two sliding windows on the event log, two feature vectors are derived from the two windows trace variants, these windows are named reference and detection windows. Then variants of the two windows are compared by applying statistical G-test and finally the drifts are identified.  In statistics, G-test is likelihood-ratio or maximum likelihood statistical significance test. Experiments on artificial databases show the correctness of the method and its superiority to the previous methods. In the proposed method, the detection accuracy is 0.06% better than state-of-the-art methods on average http://jsdp.rcisp.ac.ir/article-1-1051-en.pdf Concept drift event log process mining business processes variant per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 87 100 article Mir Mohammad Alipour alipour@ubonab.ac.ir 1 Mohsen Abdolhosseinzadeh mohsen.ab@ubonab.ac.ir 2 University of Bonab University of Bonab Recent researches show that diverse systems in many different areas can be represented as complex networks. Examples of these include the Internet, social networks and so on. In each case, the system can be modeled as a complex and very large network consisting of a large number of entities and associations between them. Most of these networks are generally sparse in global yet dense in local. They have vertices in a group structure and the vertices within a group have higher density of edges while vertices among groups have lower density of edges. Such a structure is called community and is one of the important features of the network and is able to reveal many hidden characteristics of the networks. Today, community detection is used to improve the efficiency of search engines and discovery of terrorist organizations on the World Wide Web. Community detection is a challenging NP-hard optimization problem that consists of searching for communities. It is assumed that the nodes of the same community share some properties that enable the detection of new characteristics or functional relationships in a network. Although there are many algorithms developed for community detection, most of them are unsuitable when dealing with large networks due to their computational cost. Nowadays, multiagent systems have been used to solve different problems, such as constraint satisfaction problems and combinatorial optimization problems with satisfactory results. In this paper, a new multiagent reinforcement learning algorithm is proposed for community detection in complex networks. Each agent in the multiagent system is an autonomous entity with different learning parameters. Based on the cooperation among the learning agents and updating the action probabilities of each agent, the algorithm interactively will identify a set of communities in the input network that are more densely connected than other communities. In other words, some independent agents interactively attempt to identify communities and evaluate the quality of the communities found at each stage by the normalized cut as objective function; then, the probability vectors of the agents are updated based on the results of the evaluation. If the quality of the community found by an agent in each of the stages is better than all the results produced so far, then it is referred to as the successful agent and the other agents will update their probability vectors based on the result of the successful agent. In the experiments, the performance of the proposed algorithm is validated on four real-world benchmark networks: the Karate club network, Dolphins network, Political books network and College football network, and synthetic LFR benchmark graphs with scales of 1000 and 5000 nodes. LFR networks are suitable for systematically measuring the property of an algorithm. Experimental results show that proposed approach has a good performance and is able to find suitable communities in large and small scale networks and is capable of detecting the community in complex networks In terms of speed, precision and stability. Moreover, according to the systematic comparison of the results obtained by the proposed algorithm with four state-of-the-art community detection algorithms, our algorithm outperforms the these algorithms in terms of modularity and NMI; also, it can detect communities in small and large scale networks with high speed, accuracy, and stability, where it is capable of managing large-scale networks up to 5000 nodes. http://jsdp.rcisp.ac.ir/article-1-1084-en.pdf Complex networks Community detection Multiagent systems Reinforcement learning Modularity Q per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 101 110 article mohammad jafarabad tcsms@yahoo.com 1 Rouhollah Dianat rouhollahdianat@gmail.com 2 University of Qom University of Qom For data mining studies, due to the complexity of doing feature selection process in tasks by hand, we need to send some of labeling to the workers with crowdsourcing activities. The process of outsourcing data mining tasks to users is often handled by software systems without enough knowledge of the age or geography of the users' residence. We use convolutional neural network, for doing classification in six classes: USAGE, TOPIC, COMPARE, MODEL-FEATURE, RESULT and PART-WHOLE. This article extracts the data from the abstract of 450 scientific articles and it is a total of 835 relations. One hundred of these abstracts have been selected by the crowdsourcing. Classification results in this article have been done with a slight improvement in accuracy. In this study, we computed the classification results on a combination of vocabulary vectors with using of 450 abstract relation data (100 crowd source datasets with 350 standards). The results of the implementation of the classification algorithm give us performance improvement. This paper uses the population power to perform preparing data mining works. The proposed method by adding crowdsource data to the previous data was able to obtain better results rather than the top 5 methods. http://jsdp.rcisp.ac.ir/article-1-1074-en.pdf Glove Word2vec Crowdsourcing word embedding classification per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 111 124 article Masoumeh Rezaei mrezaei@ece.usb.ac.ir 1 Mehdi Rezaeian mrezaeian@yazd.ac.ir 2 Vali Derhami vderhami@yazd.ac.ir 3 Sistan and Baluchestan University Yazd University Yazd University The processing of point clouds is one of the growing areas in machine vision. With the advent of inexpensive depth sensors, there has been a great interest in point clouds to detect three-dimensional objects. In general, 3D object recognition methods are alienated into two classes: local and global feature-based methods.  In global feature-based methods, the entire shape of the model is described, while in local methods, the geometric properties of the local area around a point are used to obtain the characteristic of the point. Unlike global methods, local methods do not entail any segmentation and they are more robust to clutter and occlusion. The local feature-based methods extract some geometric features from local surfaces around specific points named keypoints. The geometric features of a keypoint are encoded into a feature descriptor. How to describe the environment around a keypoint is the main challenge of these methods. The commonly used local feature-based methods often are sensitive to noise, varying mesh resolution, and rigid transformation. To overcome such disadvantages, in this paper, a new local feature descriptor based on the Mercator projection is proposed. The Mercator projection is one of the most popular 3D to 2D projections that can preserve true distance, direction, and relative longitude and latitude between any two points in point clouds. To evaluate, the proposed method has been compared with several state-of-the-art descriptor methods. The superiority of this method over other methods is shown by using the criteria of square Root Mean Square Error (RMSE), Recall versus 1-Precision Curve (RPC), and registration correction, rotation, and translation errors, and it is proved that this method has good descriptiveness power and it is robust to noise and varying mesh resolution. Introduction In this paper, we propose a new local descriptor to provide robust and precise geometric features. The geometric features are extracted using the Mercator projection of the neighborhood sphere. Our contributions are as follows: (1) The proposed descriptor directly learns from the point clouds (2) using the proposed method, there is only one representation for each point so the problem of multiple representations of a point is addressed. Also, the Mercator projection has many properties that make it appropriate for data representations in a point cloud. (3) It can accurately describe the geometric properties around a point. (3) The Mercator projection is a conformal projection so it preserves true distances, directions, and relative longitudes and latitudes. (4) It keeps small element geometry, which means Mercator projection preserves the shapes of small regions.  The proposed method Given a query point p, a sphere of radius r is centered at p for determining the neighbor points. Then Mercator projection is used for mapping the sphere into a plane with considering the Local reference frame (LRF) as previously suggested by Tombaret al. (2010b). The Mercator projection is a cylindrical projection that was proposed by G. Mercator in 1569. In this projection, the surface of a sphere is mapped into a plane. It preserves true distances, directions, and relative longitudes and latitudes. The Mercator projection for each point is identified using two following equations: (2) where λ is the  longitude and φ is the  latitude of a point  in the sphere, and (x,  y) represents corresponding point  in the Cartesian map. For extracting images as the input of the Siamese network, we need ranges for achieved x and  y. The variable x is in the interval [−π,  π] but range of y is different for the Mercator projection of each keypoint. As a result, the minimum and maximum of the variable y for all neighbor points are considered as the range of y, then a histogram 30 × 30 is measured. The Mercator projections of all neighbors are defined and the number of points  in each bin counted. Then we normalize the histogram by dividing each bin by the total number of neighbor points, it causes more robustness to noise and mesh resolution. Results and discussion The performance of the proposed method is evaluated on the Bologna (Tombari et al., 2010c) and John Burkardt in terms of RMSE, RPC and registration correction rate, rotation and translation errors. The proposed outperforms other methods in term of RPC also the results show that the method is robust to noise, rigid transformation and varying mesh resolution. http://jsdp.rcisp.ac.ir/article-1-1163-en.pdf Point cloud 3D object recognition Local descriptor Mercator projection per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 125 136 article Davood Zabihzadeh d.zabihzadeh@gmail.com 1 Saeed Zahedi s.zahedi@mail.um.ac.ir 2 Reza Monsefi monsefi@um.ac.ir 3 Sabzevar University of New Technology Ferdowsi University of Mashhad Ferdowsi University of Mashhad Many algorithms in machine learning, pattern recognition, and data mining are based on a similarity/distance measure. For instance, the kNN classifier and clustering algorithms such as k-means require a similarity/distance function. Also, in Content-Based Information Retrieval (CBIR) systems, we need to rank the retrieved objects based on the similarity to the query. As generic measures like Euclidean and cosine similarity are not appropriate in many applications, metric learning algorithms have been developed with the aim of learning an optimal distance function from data. These methods often need training data in the form of pair or triplet sets. Nowadays, this training data is popularly obtained via crowdsourcing from the Internet.  Therefore, this information may be contaminated with label noise resulting in the poor performance of the learned metric. In some datasets, even it is possible that the learned metrics perform worse than the general ones such as Euclidean. To address this emerging challenge, we present a new robust metric learning algorithm that can identify outliers and label noise simultaneously from training side information. For this purpose, we model the probability distribution of label noise based on information in the training data. The proposed distribution function efficiently assigns the high probability to the data points contaminated with label noise. On the other hand, its value on the normal instances is near zero. Afterward, we weight the training instances according to these probabilities in our metric learning optimization problem. The proposed optimization problem can be solved using available SVM libraries such as LibSVM efficiently. Note that the proposed approach for identifying data with label noise is general and can easily be applied to any existing metric learning algorithms. After the metric learning phase, we utilized both the weights and the learned metric to enhance the accuracy of the metric-based classifier such as kNN. Several experiments are conducted on both real and synthetic datasets. The results confirm that the proposed algorithm enhances the performance of the learned metric in the presence of label noise and considerably outperforms state-of-the-art peer methods at different noise levels. http://jsdp.rcisp.ac.ir/article-1-1096-en.pdf Robust Metric Learning Label Noise Outlier Distance Measure per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 137 152 article Zahra Abedi zahra.abedi@stu.yazd.ac.ir 1 Mahdi Yazdian-Dehkordi yazdian@yazd.ac.ir 2 Yazd University Yazd University Synthetic Aperture Radar (SAR) is widely used in different weather conditions for various applications such as mapping, remote sensing, urban, civil, and military monitoring. Recently, a new radar sensor called Video SAR (ViSAR) has been developed to capture sequential frames from moving objects for environmental monitoring applications such as image or video segmentation, classification and change detection. Same as SAR images, the major problem of ViSAR is the presence of speckle noise. In this paper, the performance of several image-based denoising methods is studied for de-speckling of ViSAR frames through “Frame-by-Frame”, “Averaging” and “3D” schemes. In “Frame-by-Frame” scheme, each video frame is denoised independently of the other frames; whereas, in “Averaging” scheme, the denoised images are averaged along a time window. In “3D” scheme, denoising is performed on 3D blocks in space-time (x-y-t) domain. In addition to these schemes, a novel extension on SAR-BM3D method, called ViSAR Incremental BM3D (ViSAR-IBM3D) approach is proposed for video denoising. The SAR-BM3D method performs denoising in two steps. At the first step, it uses wavelet denoising to primitively denoise the original image; in the next step, this image in combination with the original image are used to estimate the final denoised image. The main challenge of SAR-BM3D method is high time complexity especially for video frames. Here, in ViSAR-IBM3D, we benefit from the correlation between the frames of video and utilize the denoised images in previous frame to de-speckle the current frame. The proposed method can remarkably reduce the time complexity and improve preserving the details and the contrast of the denoised frames. The experimental results evaluated on real-world ViSAR video as well as video with simulated noises show that the proposed 3D filtering scheme and the proposed ViSAR-IBM3D method achieve better denoising performance than the other ones. http://jsdp.rcisp.ac.ir/article-1-1044-en.pdf SAR ViSAR Noise Speckle SAR-BM3D ViSAR-IBM3D per Research Center on Developing Advanced Technologies Signal and Data Processing 2538-4201 2538-421X 2022-05 19 1 153 166 article jaafar partabian Jaafar_partabian@yahoo.com 1 vahid rafe v-rafe@araku.ac.ir 2 hamid parvin parvinhamid@gmail.com 3 samad nejatian samad.nej.2007@gmail.com 4 Karamollah bagherifard k.bagheri@iauyasooj.ac.ir 5 azad university arak university azad university azad university Yasooj Branch, Islamic Azad University The model checking technique is a formal and effective method for verifying software systems, which analyses it via generating and examining all possible states of a model of the software system. In safety-critical systems, one could not admit the risk of error even in the testing process, therefore it is necessary to carry out the verification process before implementation and at the model level. Using this technique to evaluate properties such as security entails all available states (all state space) being generated, then the state space of the system in question be carefully examined. The main challenge of the model checking technique in large and complex systems with wide or infinite state space is the problem of state space explosion (lack of memory in the generation of all possible states). Graph transformation systems are one of the most widely used formal modeling systems and a suitable solution for modeling and checking complex systems. In systems where security property verification is not possible, the security feature can be refuted by searching for an accessible mode in which a specific configuration (e.g. error or undesirable behavior) occurs. Recent studies advocate that partial and intelligent exploration of part of the state space could be a good solution to the problem of state space explosion. The goal of this study is to use the random forest algorithm in the model checking which can solve the problem of state space explosion by selecting a few promising paths. A path is hopeful whenever the probability of reaching an answer through this path is higher than other paths. In the proposed method, a small model of the system is first created using the official language of the Graph Description System (GTS). Afterwards, a training data set of paths to the goal is generated from the small model mode space. The generated training data set is then provided to the random forest algorithm to identify and discover the logical relationships within it. In the next stage, the acquired knowledge is used to intelligently explore the incomplete space of the large model state. The proposed approach is used in the verification of the reachability property and to refute the safety feature in large and complex systems where it is impossible to generate the entire system state space. In order to evaluate the proposed approach, it has been implemented in GROOVE which is an open source tool for designing and checking models in graph conversion systems. The results indicate that the proposed method performs better than the compared methods in terms of average running time and the length of the generated witness. http://jsdp.rcisp.ac.ir/article-1-1075-en.pdf Software systems verification Knowledge discovery State space explosion intelligent search