Yahya Dorostkar Navaei
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yahya Dorostkar Navaei has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Sajad Einy
Computer Engineering Department
6 shared publications - Cemil Öz
Sakarya University
3 shared publications - Hasan Saygin
Application and Research Center for Advanced Studies Istanbul Aydin University Istanbul Türkiye
2 shared publications - Junde Chen
Earth Systems Science and Data Solutions Lab, Chapman University, Orange, CA, USA
2 shared publications - Yaser Ahangari Nanehkaran
School of Information Engineering, Yancheng Teachers University, Yancheng, China
2 shared publications - Zhu Licai
School of Information Engineering
2 shared publications
Latest Publications
- Open Access
- 14 June 2023
- Published by Hindawi Limited in Wireless Communications and Mobile Computing
- Vol. 2023, 1-18
A Clustering-Based Routing Protocol Using Path Pattern Discovery Method to Minimize Delay in VANET
In vehicular ad hoc networks (VANETs), vehicle-to-vehicle (V2V) communications can link vehicles to each other, and vehicle-to-infrastructure (V2I) messaging and communications can link roadside infrastructure such as routers. The vehicles in these networks act as relays that transmit critical messages in the network. Due to the high-speed movement of vehicles on the road, real-time messaging and minimizing the delay in sending messages is one of the most important objectives of VANET developers. On the other hand, the high mobility of vehicles causes communication interruptions and decreases the data delivery rate in VANET. To overcome this issue, predicting the path of vehicles can play an important role in sending data from the source to the destination. When an accident occurs on the road, the messages that are sensed by the imbedded sensors in the vehicles need to be sent, and if they are sent by the vehicles that change their route, these messages will not be sent to the destination and the performance of the network will be disturbed. Previous methods in the literature for data transmission in intervehicular networks have focused more on reliability and trust, and little attention has been paid to the prediction of vehicle movement paths in these types of networks. Therefore, for fast and reliable data transmission in VANET, accurate prediction of vehicle movement and creation of movement patterns can be effective in message transmission delay and data delivery rate. In this paper, we present an approach using a combination of cluster-based routing protocols and pattern discovery methods to minimize latency in VANETs. The outline of the proposed method has four modules: primary data collection and analysis, primary data preparation and analysis, pattern extraction and vehicle route discovery, and vehicle clustering and data/information transmission routing. The simulation results show that the proposed method with a delivery rate of 88.56% has significantly improved compared to the previous methods in terms of package delivery rate. Also, the proposed method with a total delay of 24.566 ms has a shorter delay than the previous methods in terms of message sending delay in the network.
- 14 June 2023
- Published by Hindawi Limited in Wireless Communications and Mobile Computing
- Vol. 2023, 1-18
Open AccessA Clustering-Based Routing Protocol Using Path Pattern Discovery Method to Minimize Delay in VANET
In vehicular ad hoc networks (VANETs), vehicle-to-vehicle (V2V) communications can link vehicles to each other, and vehicle-to-infrastructure (V2I) messaging and communications can link roadside infrastructure such as routers. The vehicles in these networks act as relays that transmit critical messages in the network. Due to the high-speed movement of vehicles on the road, real-time messaging and minimizing the delay in sending messages is one of the most important objectives of VANET developers. On the other hand, the high mobility of vehicles causes communication interruptions and decreases the data delivery rate in VANET. To overcome this issue, predicting the path of vehicles can play an important role in sending data from the source to the destination. When an accident occurs on the road, the messages that are sensed by the imbedded sensors in the vehicles need to be sent, and if they are sent by the vehicles that change their route, these messages will not be sent to the destination and the performance of the network will be disturbed. Previous methods in the literature for data transmission in intervehicular networks have focused more on reliability and trust, and little attention has been paid to the prediction of vehicle movement paths in these types of networks. Therefore, for fast and reliable data transmission in VANET, accurate prediction of vehicle movement and creation of movement patterns can be effective in message transmission delay and data delivery rate. In this paper, we present an approach using a combination of cluster-based routing protocols and pattern discovery methods to minimize latency in VANETs. The outline of the proposed method has four modules: primary data collection and analysis, primary data preparation and analysis, pattern extraction and vehicle route discovery, and vehicle clustering and data/information transmission routing. The simulation results show that the proposed method with a delivery rate of 88.56% has significantly improved compared to the previous methods in terms of package delivery rate. Also, the proposed method with a total delay of 24.566 ms has a shorter delay than the previous methods in terms of message sending delay in the network.
- Open Access
- 17 April 2023
- Published by Wiley in Scientific Programming
- Vol. 2023, 1-11
Local Binary Convolutional Neural Networks' Long Short-Term Memory Model for Human Embryos' Anomaly Detection
Accurate selection of embryos with the maximum implementation condition is a necessary step to increase the effectiveness of fertility treatment in in vitro fertilization (IVF). The deep learning algorithms presented high potential for monitoring and visualizing embryo features such as cell numbers and their morphological development in time series manner. Due to the ability of the computer vision and deep learning algorithms, this paper aimed to present a novel deep learning approach to distinguish simultaneous abnormality of embryos in time-lapse systems for detecting live and non-live births in IVF. The approach is composed of local binary convolutional neural network (LBCNN) and long short-term memory (LSTM). The LBCNN improved accuracy of classification by employing deep and local feature sets with lower number of learnable parameters in comparison with a standard convolutional layer. Moreover, LSTM network is employed to analyze temporal information of time-lapse embryos. The results indicate that the proposed approach achieves significant results in ROC analysis (0.98) in 5days of monitoring compared to state-of-the-art models. In addition, the approach showed compatible results in early diagnosis of abnormality detection (72hours) with 82.8 accuracy of classification compared to the pretrained well-known convolutional neural network (CNN) models and baseline CNN.
- 17 April 2023
- Published by Wiley in Scientific Programming
- Vol. 2023, 1-11
Open AccessLocal Binary Convolutional Neural Networks' Long Short-Term Memory Model for Human Embryos' Anomaly Detection
Accurate selection of embryos with the maximum implementation condition is a necessary step to increase the effectiveness of fertility treatment in in vitro fertilization (IVF). The deep learning algorithms presented high potential for monitoring and visualizing embryo features such as cell numbers and their morphological development in time series manner. Due to the ability of the computer vision and deep learning algorithms, this paper aimed to present a novel deep learning approach to distinguish simultaneous abnormality of embryos in time-lapse systems for detecting live and non-live births in IVF. The approach is composed of local binary convolutional neural network (LBCNN) and long short-term memory (LSTM). The LBCNN improved accuracy of classification by employing deep and local feature sets with lower number of learnable parameters in comparison with a standard convolutional layer. Moreover, LSTM network is employed to analyze temporal information of time-lapse embryos. The results indicate that the proposed approach achieves significant results in ROC analysis (0.98) in 5days of monitoring compared to state-of-the-art models. In addition, the approach showed compatible results in early diagnosis of abnormality detection (72hours) with 82.8 accuracy of classification compared to the pretrained well-known convolutional neural network (CNN) models and baseline CNN.
- Open Access
- 1 July 2022
- Published by Hindawi Limited in Wireless Communications and Mobile Computing
- Vol. 2022, 1-22
Load Balancing in Edge Computing Using Integer Linear Programming Based Genetic Algorithm and Multilevel Control Approach
Due to the proliferation of requests in heterogeneous resources in edge computing, the existence of a large number of tasks and workloads in virtual machines in the edge computing environment is inevitable. Thus, load balancing strives to facilitate an even distribution of workload across available resources. Its purpose is to provide continuous service and to ensure fair load distribution among resources. Load balancing, with the aim of minimizing response time for tasks and improving resource efficiency, tries to do the proper mapping of tasks among virtual machines at a lower cost. Flow scheduling, on the other hand, assigns a task (group of tasks) to computational resources by prioritizing tasks, so that the relationship between them is maintained. Therefore, in this research, a hierarchical control framework for load balancing and assignment of tasks in edge computing services is presented in order to create load balancing. In the proposed method, in the first level, the genetic algorithm receives a set of tasks in a workflow. Genetic algorithm prioritizes and assigns tasks to resources according to time constraints, resource processing power, resource availability, and task cost. For this purpose, the integer linear programming optimization in the evaluation function of the genetic algorithm will be utilized. In the second level, considering the past load distribution in edge resources, we estimate the probability of load distribution among sources according to hidden Markov model (HMM). Finally, in order to optimally map tasks to the virtual machines in each host, we will use game theory with service quality factors as an evaluation function. Previous methods have provided a hierarchical control framework that aims to achieve conflicting goals within a data center, but does not use linear programming. Considering the use of service quality criteria as evaluation function parameters in heuristic and optimization methods in this research, it is expected that the results of this research will improve compared to previous methods.
- 1 July 2022
- Published by Hindawi Limited in Wireless Communications and Mobile Computing
- Vol. 2022, 1-22
Open AccessLoad Balancing in Edge Computing Using Integer Linear Programming Based Genetic Algorithm and Multilevel Control Approach
Due to the proliferation of requests in heterogeneous resources in edge computing, the existence of a large number of tasks and workloads in virtual machines in the edge computing environment is inevitable. Thus, load balancing strives to facilitate an even distribution of workload across available resources. Its purpose is to provide continuous service and to ensure fair load distribution among resources. Load balancing, with the aim of minimizing response time for tasks and improving resource efficiency, tries to do the proper mapping of tasks among virtual machines at a lower cost. Flow scheduling, on the other hand, assigns a task (group of tasks) to computational resources by prioritizing tasks, so that the relationship between them is maintained. Therefore, in this research, a hierarchical control framework for load balancing and assignment of tasks in edge computing services is presented in order to create load balancing. In the proposed method, in the first level, the genetic algorithm receives a set of tasks in a workflow. Genetic algorithm prioritizes and assigns tasks to resources according to time constraints, resource processing power, resource availability, and task cost. For this purpose, the integer linear programming optimization in the evaluation function of the genetic algorithm will be utilized. In the second level, considering the past load distribution in edge resources, we estimate the probability of load distribution among sources according to hidden Markov model (HMM). Finally, in order to optimally map tasks to the virtual machines in each host, we will use game theory with service quality factors as an evaluation function. Previous methods have provided a hierarchical control framework that aims to achieve conflicting goals within a data center, but does not use linear programming. Considering the use of service quality criteria as evaluation function parameters in heuristic and optimization methods in this research, it is expected that the results of this research will improve compared to previous methods.
- Open Access
- 1 January 2022
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2022 (1)
Diagnosis of Chronic Diseases Based on Patients’ Health Records in IoT Healthcare Using the Recommender System
Due to the growth of IoT applications, especially health care, the information of patients’ health records using data collection from IoT‐connected devices has been considered. Biological data of patients in the health record helps to monitor the patient’s status and identify various diseases. Chronic diseases are a type of silent disease that, if not diagnosed in time, can cause irreparable damage to patients. The use of patients’ medical record data for early diagnosis of chronic diseases has recently attracted the attention of many researchers. On the other hand, the application of machine learning methods in the form of recommender systems has taken an important step in improving medical services and health care. In this paper, a medical recommender system was presented to identify and treat chronic diseases using an IoT device. In the present method, the electronic patient health record dataset that is loaded in the PhysioNet data repository has been used. In the present dataset, patients’ health records have been recorded according to the identified diseases and the physician’s diagnosis. In the proposed method, the K‐nearest neighbor classification method is used to identify the type of disease, and the collaborative filtering method is used to find the appropriate treatment for patients. The results of the implementation of the proposed method show that this approach, based on the use of symptom similarity among patients, has good accuracy in diagnosing and predicting chronic diseases and has provided higher results than previous methods.
- 1 January 2022
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2022 (1)
Open AccessDiagnosis of Chronic Diseases Based on Patients’ Health Records in IoT Healthcare Using the Recommender System
Due to the growth of IoT applications, especially health care, the information of patients’ health records using data collection from IoT‐connected devices has been considered. Biological data of patients in the health record helps to monitor the patient’s status and identify various diseases. Chronic diseases are a type of silent disease that, if not diagnosed in time, can cause irreparable damage to patients. The use of patients’ medical record data for early diagnosis of chronic diseases has recently attracted the attention of many researchers. On the other hand, the application of machine learning methods in the form of recommender systems has taken an important step in improving medical services and health care. In this paper, a medical recommender system was presented to identify and treat chronic diseases using an IoT device. In the present method, the electronic patient health record dataset that is loaded in the PhysioNet data repository has been used. In the present dataset, patients’ health records have been recorded according to the identified diseases and the physician’s diagnosis. In the proposed method, the K‐nearest neighbor classification method is used to identify the type of disease, and the collaborative filtering method is used to find the appropriate treatment for patients. The results of the implementation of the proposed method show that this approach, based on the use of symptom similarity among patients, has good accuracy in diagnosing and predicting chronic diseases and has provided higher results than previous methods.
- Open Access
- 1 January 2022
- Published by Wiley in Complexity
- Vol. 2022 (1)
Local and Deep Features Based Convolutional Neural Network Frameworks for Brain MRI Anomaly Detection
A brain tumor is an abnormal mass or growth of a cell that leads to certain death, and this is still a challenging task in clinical practice. Early and correct diagnosis of this type of cancer is very important for the treatment process. For this reason, this study aimed to develop computer‐aided systems for the diagnosis of brain tumors. In this research, we proposed three different end‐to‐end deep learning approaches for analyzing effects of local and deep features for brain MRI images anomaly detection. The first proposed system is Directional Bit‐Planes Deep Autoencoder (DBP‐DAE) which extracts and learns local and direction features. The DBP‐DAE by decomposition of a local binary pattern (LBP) into eight bit‐planes extracts are directional and inherent local‐structure features from the input image and learns robust feature for classification purposes. The second one is a Dilated Separable Residual Convolutional Network (DSRCN) which extracts high (deep) and low‐level features. The main advantage of this approach is that it is robust and shows stable results regardless to size of image database and to solve overfitting problems. To explore the effects of mixture of local and deep extracted feature on accuracy of classification of brain anomaly, a multibranch convolutional neural network approach is proposed. This approach is designed according to combination of DBP‐DAE and DSRCN in an end‐to‐end manner. Extensive experiments conducted based on brain tumor in MRI image public access databases and achieves significant results compared to state‐of‐the‐art algorithms. In addition, we discussed the effectiveness and applicability of CNNs with a variety of different features and architectures for brain abnormalities such as Alzheimer’s.
- 1 January 2022
- Published by Wiley in Complexity
- Vol. 2022 (1)
Open AccessLocal and Deep Features Based Convolutional Neural Network Frameworks for Brain MRI Anomaly Detection
A brain tumor is an abnormal mass or growth of a cell that leads to certain death, and this is still a challenging task in clinical practice. Early and correct diagnosis of this type of cancer is very important for the treatment process. For this reason, this study aimed to develop computer‐aided systems for the diagnosis of brain tumors. In this research, we proposed three different end‐to‐end deep learning approaches for analyzing effects of local and deep features for brain MRI images anomaly detection. The first proposed system is Directional Bit‐Planes Deep Autoencoder (DBP‐DAE) which extracts and learns local and direction features. The DBP‐DAE by decomposition of a local binary pattern (LBP) into eight bit‐planes extracts are directional and inherent local‐structure features from the input image and learns robust feature for classification purposes. The second one is a Dilated Separable Residual Convolutional Network (DSRCN) which extracts high (deep) and low‐level features. The main advantage of this approach is that it is robust and shows stable results regardless to size of image database and to solve overfitting problems. To explore the effects of mixture of local and deep extracted feature on accuracy of classification of brain anomaly, a multibranch convolutional neural network approach is proposed. This approach is designed according to combination of DBP‐DAE and DSRCN in an end‐to‐end manner. Extensive experiments conducted based on brain tumor in MRI image public access databases and achieves significant results compared to state‐of‐the‐art algorithms. In addition, we discussed the effectiveness and applicability of CNNs with a variety of different features and architectures for brain abnormalities such as Alzheimer’s.
- Open Access
- 1 January 2022
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2022 (1)
A Novel Energy‐Aware Routing in Wireless Sensor Network Using Clustering Based on Combination of Multiobjective Genetic and Cuckoo Search Algorithm
The development of various applications of wireless sensor networks has aroused great interest in using these types of networks in various fields. These networks, without infrastructure and self‐organization, are easily deployed in most environments and collect information about environmental phenomena for analysis and proper response to accidents and send them to the basic centers. They do. Wireless sensor networks are made up of some sensor nodes that both act as sensors and act as relay nodes concerning to each other. On the other hand, the lack of infrastructure in these networks has limited resources so that the nodes of the battery are fed with limited energy. Due to the location of networks in difficult and impassable areas, it is not possible to recharge or replace the node battery. Therefore, saving energy consumption in this type of network is one of the most important challenges. Since the rate of energy consumption when sensing information and receiving data packets from another node is a fixed value, so sensor nodes have the highest energy consumption when sending data. Therefore, routing methods try to reduce energy consumption based on systematic approaches. One of the most promising solutions to reduce energy consumption in wireless sensor networks is to cluster the nodes and select the threaded node based on the data transfer parameters so that the average energy consumption in the nodes is reduced and the network lifetime is increased. Therefore, in this research, a new optimization approach using multiobjective genetic algorithm and cuckoo algorithm for clustering wireless sensor networks is presented. In this study, in order to select clustered nodes from a multiobjective genetic algorithms based on reducing intracluster distances and reducing energy consumption in cluster member nodes and near‐optimal routing based on cuckoo optimization algorithm to transfer information between nodes have been used in the direction of the cavity. The implementation results show that considering the evolutionary capabilities of the multiobjective genetic algorithm and the cuckoo optimization algorithm, the proposed method in terms of energy consumption, efficiency, delivery rate, and packet transmission latency, compared to previous methods, has improved.
- 1 January 2022
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2022 (1)
Open AccessA Novel Energy‐Aware Routing in Wireless Sensor Network Using Clustering Based on Combination of Multiobjective Genetic and Cuckoo Search Algorithm
The development of various applications of wireless sensor networks has aroused great interest in using these types of networks in various fields. These networks, without infrastructure and self‐organization, are easily deployed in most environments and collect information about environmental phenomena for analysis and proper response to accidents and send them to the basic centers. They do. Wireless sensor networks are made up of some sensor nodes that both act as sensors and act as relay nodes concerning to each other. On the other hand, the lack of infrastructure in these networks has limited resources so that the nodes of the battery are fed with limited energy. Due to the location of networks in difficult and impassable areas, it is not possible to recharge or replace the node battery. Therefore, saving energy consumption in this type of network is one of the most important challenges. Since the rate of energy consumption when sensing information and receiving data packets from another node is a fixed value, so sensor nodes have the highest energy consumption when sending data. Therefore, routing methods try to reduce energy consumption based on systematic approaches. One of the most promising solutions to reduce energy consumption in wireless sensor networks is to cluster the nodes and select the threaded node based on the data transfer parameters so that the average energy consumption in the nodes is reduced and the network lifetime is increased. Therefore, in this research, a new optimization approach using multiobjective genetic algorithm and cuckoo algorithm for clustering wireless sensor networks is presented. In this study, in order to select clustered nodes from a multiobjective genetic algorithms based on reducing intracluster distances and reducing energy consumption in cluster member nodes and near‐optimal routing based on cuckoo optimization algorithm to transfer information between nodes have been used in the direction of the cavity. The implementation results show that considering the evolutionary capabilities of the multiobjective genetic algorithm and the cuckoo optimization algorithm, the proposed method in terms of energy consumption, efficiency, delivery rate, and packet transmission latency, compared to previous methods, has improved.
- Open Access
- 1 January 2022
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2022 (1)
Anomaly Detection in Heart Disease Using a Density‐Based Unsupervised Approach
Cardiovascular disease is one of the most common diseases in the modern world, which, if diagnosed early, can greatly reduce the damage to the patient. Diagnosis of heart disease requires great care, and in some cases, the process can be disrupted by human error. Machine learning methods, especially data mining, have gained international acceptance in almost all aspects of life, especially the prediction of heart disease. On the other hand, datasets related to heart patients have many biological features that most of these features do not have a direct impact on diagnosis. By removing redundant features from the dataset, in addition to reducing computational complexity, the accuracy of heart patients’ predictions can also be increased. This paper presents a density‐based unsupervised approach to the diagnosis of abnormalities in heart patients. In this method, the basic features in the dataset are first selected based on the filter‐based feature selection approach. Then, the DBSCAN clustering method with adaptive parameters has used to increase the clustering accuracy of healthy instances and to determine abnormal instances as cardiac patients. Partition clustering methods suffer from the selection of the number of clusters and the initial central points and are very sensitive to noise. The DBSCAN method solves these problems by creating density‐based clusters, but the selection of the neighborhood radius threshold and the number of connected points in the neighborhood remains unresolved. In the proposed method, these two parameters are selected adaptively to achieve the highest accuracy for the diagnosis and prediction of heart patients. The results of the experiments show that the accuracy of the proposed method for predicting heart patients is approximately 95%, which has improved in comparison with previous methods.
- 1 January 2022
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2022 (1)
Open AccessAnomaly Detection in Heart Disease Using a Density‐Based Unsupervised Approach
Cardiovascular disease is one of the most common diseases in the modern world, which, if diagnosed early, can greatly reduce the damage to the patient. Diagnosis of heart disease requires great care, and in some cases, the process can be disrupted by human error. Machine learning methods, especially data mining, have gained international acceptance in almost all aspects of life, especially the prediction of heart disease. On the other hand, datasets related to heart patients have many biological features that most of these features do not have a direct impact on diagnosis. By removing redundant features from the dataset, in addition to reducing computational complexity, the accuracy of heart patients’ predictions can also be increased. This paper presents a density‐based unsupervised approach to the diagnosis of abnormalities in heart patients. In this method, the basic features in the dataset are first selected based on the filter‐based feature selection approach. Then, the DBSCAN clustering method with adaptive parameters has used to increase the clustering accuracy of healthy instances and to determine abnormal instances as cardiac patients. Partition clustering methods suffer from the selection of the number of clusters and the initial central points and are very sensitive to noise. The DBSCAN method solves these problems by creating density‐based clusters, but the selection of the neighborhood radius threshold and the number of connected points in the neighborhood remains unresolved. In the proposed method, these two parameters are selected adaptively to achieve the highest accuracy for the diagnosis and prediction of heart patients. The results of the experiments show that the accuracy of the proposed method for predicting heart patients is approximately 95%, which has improved in comparison with previous methods.
- Open Access
- 12 March 2021
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2021, 1-10
The Anomaly- and Signature-Based IDS for Network Security Using Hybrid Inference Systems
With the expansion of communication in today’s world and the possibility of creating interactions between people through communication networks regardless of the distance dimension, the issue of creating security for the data and information exchanged has received much attention from researchers. Various methods have been proposed for this purpose; one of the most important methods is intrusion detection systems to quickly detect intrusions into the network and inform the manager or responsible people to carry out an operational set to reduce the amount of damage caused by these intruders. The main challenge of the proposed intrusion detection systems is the number of erroneous warning messages generated and the low percentage of accurate detection of intrusions in them. In this research, the Suricata IDS/IPS is deployed along with the NN model for the metaheuristic’s manual detection of malicious traffic in the targeted network. For the metaheuristic-based feature selection, the neural network, and the anomaly-based detection, the fuzzy logic is used in this research paper. The latest stable version of Kali Linux 2020.3 is used as an attacking system for web applications and different types of operating systems. The proposed method has achieved 96.111% accuracy for detecting network intrusion.
- 12 March 2021
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2021, 1-10
Open AccessThe Anomaly- and Signature-Based IDS for Network Security Using Hybrid Inference Systems
With the expansion of communication in today’s world and the possibility of creating interactions between people through communication networks regardless of the distance dimension, the issue of creating security for the data and information exchanged has received much attention from researchers. Various methods have been proposed for this purpose; one of the most important methods is intrusion detection systems to quickly detect intrusions into the network and inform the manager or responsible people to carry out an operational set to reduce the amount of damage caused by these intruders. The main challenge of the proposed intrusion detection systems is the number of erroneous warning messages generated and the low percentage of accurate detection of intrusions in them. In this research, the Suricata IDS/IPS is deployed along with the NN model for the metaheuristic’s manual detection of malicious traffic in the targeted network. For the metaheuristic-based feature selection, the neural network, and the anomaly-based detection, the fuzzy logic is used in this research paper. The latest stable version of Kali Linux 2020.3 is used as an attacking system for web applications and different types of operating systems. The proposed method has achieved 96.111% accuracy for detecting network intrusion.
- Open Access
- 1 January 2021
- Published by Wiley in Journal of Sensors
- Vol. 2021 (1)
IoT Cloud‐Based Framework for Face Spoofing Detection with Deep Multicolor Feature Learning Model
A face‐based authentication system has become an important topic in various fields of IoT applications such as identity validation for social care, crime detection, ATM access, computer security, etc. However, these authentication systems are vulnerable to different attacks. Presentation attacks have become a clear threat for facial biometric‐based authentication and security applications. To address this issue, we proposed a deep learning approach for face spoofing detection systems in IoT cloud‐based environment. The deep learning approach extracted features from multicolor space to obtain more information from the input face image regarding luminance and chrominance data. These features are combined and selected by the Minimum Redundancy Maximum Relevance (mRMR) algorithm to provide an efficient and discriminate feature set. Finally, the extracted deep color‐based features of the face image are used for face spoofing detection in a cloud environment. The proposed method achieves stable results with less training data compared to conventional deep learning methods. This advantage of the proposed approach reduces the time of processing in the training phase and optimizes resource management in storing training data on the cloud. The proposed system was tested and evaluated based on two challenging public access face spoofing databases, namely, Replay‐Attack and ROSE‐Youtu. The experimental results based on these databases showed that the proposed method achieved satisfactory results compared to the state‐of‐the‐art methods based on an equal error rate (EER) of 0.2% and 3.8%, respectively, for the Replay‐Attack and ROSE‐Youtu databases.
- 1 January 2021
- Published by Wiley in Journal of Sensors
- Vol. 2021 (1)
Open AccessIoT Cloud‐Based Framework for Face Spoofing Detection with Deep Multicolor Feature Learning Model
A face‐based authentication system has become an important topic in various fields of IoT applications such as identity validation for social care, crime detection, ATM access, computer security, etc. However, these authentication systems are vulnerable to different attacks. Presentation attacks have become a clear threat for facial biometric‐based authentication and security applications. To address this issue, we proposed a deep learning approach for face spoofing detection systems in IoT cloud‐based environment. The deep learning approach extracted features from multicolor space to obtain more information from the input face image regarding luminance and chrominance data. These features are combined and selected by the Minimum Redundancy Maximum Relevance (mRMR) algorithm to provide an efficient and discriminate feature set. Finally, the extracted deep color‐based features of the face image are used for face spoofing detection in a cloud environment. The proposed method achieves stable results with less training data compared to conventional deep learning methods. This advantage of the proposed approach reduces the time of processing in the training phase and optimizes resource management in storing training data on the cloud. The proposed system was tested and evaluated based on two challenging public access face spoofing databases, namely, Replay‐Attack and ROSE‐Youtu. The experimental results based on these databases showed that the proposed method achieved satisfactory results compared to the state‐of‐the‐art methods based on an equal error rate (EER) of 0.2% and 3.8%, respectively, for the Replay‐Attack and ROSE‐Youtu databases.
- Open Access
- 1 January 2021
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2021 (1)
Network Intrusion Detection System Based on the Combination of Multiobjective Particle Swarm Algorithm‐Based Feature Selection and Fast‐Learning Network
Given the growth of wireless networks and the increase of the advantages and applications of communication networks, especially mobile ad hoc networks (MANETs), this type of network has attracted the attention of users and researchers more than before. The benefit of these types of networks in various kinds of networks and environments is that MANET does not require to hardware infrastructure to communicate and send and receive data packets within the network. It is one of the main reasons for using these MANET in various fields. On the other hand, the increased popularity of these networks has led to many challenges, one of the most important of which is network security. In this regard, a lack of regulatory and security infrastructure in MANETs has caused some problems in sending and receiving data, where intrusion in the network has been recognized as one of the most important issues. In MANETs, wireless notes act as a link between the source and destination nodes and play the role of relays and routers in the network. Therefore, malicious node penetration and the destruction of information packages become feasible. Today, intrusion detection systems (IDSs) are used as a solution to deal with the problem through remote monitoring of the performance and behaviors of nodes existing in wireless sensor networks. In addition to detecting malicious nodes in the network, IDSs can predict the behavior of malicious nodes in the future in most cases. Therefore, the present study introduced a network IDS (NIDS) entitled MOPSO‐FLN by using a combination of multiobjective particle swarm optimization algorithm‐ (MOPSO‐) based feature subset selection (FSS) and fast‐learning network (FLN). In this work, we used the KDD Cup99 and dataset to select features, train the network, and test the model. According to the simulation results, this method was able to improve the performance of the IDS in terms of evaluation criteria, compared to other previous methods, by creating a balance between the objectives of the number of representative features and training errors based on the evolutionary power of MOPSO.
- 1 January 2021
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2021 (1)
Open AccessNetwork Intrusion Detection System Based on the Combination of Multiobjective Particle Swarm Algorithm‐Based Feature Selection and Fast‐Learning Network
Given the growth of wireless networks and the increase of the advantages and applications of communication networks, especially mobile ad hoc networks (MANETs), this type of network has attracted the attention of users and researchers more than before. The benefit of these types of networks in various kinds of networks and environments is that MANET does not require to hardware infrastructure to communicate and send and receive data packets within the network. It is one of the main reasons for using these MANET in various fields. On the other hand, the increased popularity of these networks has led to many challenges, one of the most important of which is network security. In this regard, a lack of regulatory and security infrastructure in MANETs has caused some problems in sending and receiving data, where intrusion in the network has been recognized as one of the most important issues. In MANETs, wireless notes act as a link between the source and destination nodes and play the role of relays and routers in the network. Therefore, malicious node penetration and the destruction of information packages become feasible. Today, intrusion detection systems (IDSs) are used as a solution to deal with the problem through remote monitoring of the performance and behaviors of nodes existing in wireless sensor networks. In addition to detecting malicious nodes in the network, IDSs can predict the behavior of malicious nodes in the future in most cases. Therefore, the present study introduced a network IDS (NIDS) entitled MOPSO‐FLN by using a combination of multiobjective particle swarm optimization algorithm‐ (MOPSO‐) based feature subset selection (FSS) and fast‐learning network (FLN). In this work, we used the KDD Cup99 and dataset to select features, train the network, and test the model. According to the simulation results, this method was able to improve the performance of the IDS in terms of evaluation criteria, compared to other previous methods, by creating a balance between the objectives of the number of representative features and training errors based on the evolutionary power of MOPSO.