Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9924087
Ashish Kumar Lal, S. Karthikeyan
Cardiotocography (CTG) is a continuous recording of the fetal heart rate (FHR) obtained from an ultrasound transducer placed on the mother's abdomen. In common practice, obstetricians visually inspect the CTG signal to monitor the condition of the fetus's heart. This manual inspection is not reliable as it is prone to human error and biases. To overcome these short-comings, researchers had developed various AI-based diagnosis models for the automatic classification of CTG data. A few recent research had reported that neural network outperforms other machine learning models. Despite the advancements in automatic classification techniques, the adoption of these AI models has not been widespread due to the requirement for privacy of the patient record. The medical institutions are unwilling to share or publish these records, due to ethical and legal reasons. This discourages the deployment of such AI models and consequently hinders active and collaborative research work. To alleviate the privacy breach concern, we used a deep privacy-preserving CTG data classification model by adopting Differential Privacy (D P) framework. DP has widely been accepted as the gold standard of privacy guarantee. As privacy comes at an additional cost of slight downgrade in the model's performance. To mitigate this performance degradation, we have proposed a two stage binary classification which improves the model performance while maintaining the same privacy guarantee. The experimental results show that an improved performance of the proposed model with accuracy increased to 0.91 from 0.89 with E = 10 of (E,6)- Differential Privacy.
{"title":"Deep Learning Classification of Fetal Cardiotocography Data with Differential Privacy","authors":"Ashish Kumar Lal, S. Karthikeyan","doi":"10.1109/CSI54720.2022.9924087","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924087","url":null,"abstract":"Cardiotocography (CTG) is a continuous recording of the fetal heart rate (FHR) obtained from an ultrasound transducer placed on the mother's abdomen. In common practice, obstetricians visually inspect the CTG signal to monitor the condition of the fetus's heart. This manual inspection is not reliable as it is prone to human error and biases. To overcome these short-comings, researchers had developed various AI-based diagnosis models for the automatic classification of CTG data. A few recent research had reported that neural network outperforms other machine learning models. Despite the advancements in automatic classification techniques, the adoption of these AI models has not been widespread due to the requirement for privacy of the patient record. The medical institutions are unwilling to share or publish these records, due to ethical and legal reasons. This discourages the deployment of such AI models and consequently hinders active and collaborative research work. To alleviate the privacy breach concern, we used a deep privacy-preserving CTG data classification model by adopting Differential Privacy (D P) framework. DP has widely been accepted as the gold standard of privacy guarantee. As privacy comes at an additional cost of slight downgrade in the model's performance. To mitigate this performance degradation, we have proposed a two stage binary classification which improves the model performance while maintaining the same privacy guarantee. The experimental results show that an improved performance of the proposed model with accuracy increased to 0.91 from 0.89 with E = 10 of (E,6)- Differential Privacy.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126436306","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9924006
Poonam T. Agarkar, M. Chawhan, K. Kulat, P. Hajare
Growing demand in wireless Mobile Adhoc Networks due to their immense potential and effectiveness in varied applications had motivated researchers to control and optimize various parameters for efficient routing in mobile environment. The rapidly changing environment of the network due to mobility of nodes results in unpredictable topology and thus introduces primary challenge to design better and an efficient routing protocol. The network in such case lacks fixed infrastructure and common head centric design. MANET routing schemes use flooding to propagate packets in the network resulting retransmissions, collisions, and congestions which significantly degrade the performance of the network. Using GPS, knowing the geographical position of the sensor nodes, the protocol performance can be improved while reducing the number of retransmissions. The proposed work is a future extension of efficient flooding based on selective neighbours where the surrounding region limiting the transmission range is divided into eight quadrants or zones with the source at the centre called zone based selective neighbours (ZBSN). Flooding based on selective four neighbours suffers from few selection of hoping or forwarding nodes when the node density is low and does not meet the selection criteria around the source or forwarder node and may miss the chance of approaching the destination or requires large number of hops. This neighbour selection scheme uses the modified approach of the Adhoc on Demand Distance Vector (AODV) protocol to reduce the flooding of Route request (RREQ) packets, control the network overall traffic, improves link stability and the residual energy reducing the overheads by 24%.
{"title":"Zone Based Selective Neighbours to Mitigate Flooding & Reliable Routing for WSN","authors":"Poonam T. Agarkar, M. Chawhan, K. Kulat, P. Hajare","doi":"10.1109/CSI54720.2022.9924006","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924006","url":null,"abstract":"Growing demand in wireless Mobile Adhoc Networks due to their immense potential and effectiveness in varied applications had motivated researchers to control and optimize various parameters for efficient routing in mobile environment. The rapidly changing environment of the network due to mobility of nodes results in unpredictable topology and thus introduces primary challenge to design better and an efficient routing protocol. The network in such case lacks fixed infrastructure and common head centric design. MANET routing schemes use flooding to propagate packets in the network resulting retransmissions, collisions, and congestions which significantly degrade the performance of the network. Using GPS, knowing the geographical position of the sensor nodes, the protocol performance can be improved while reducing the number of retransmissions. The proposed work is a future extension of efficient flooding based on selective neighbours where the surrounding region limiting the transmission range is divided into eight quadrants or zones with the source at the centre called zone based selective neighbours (ZBSN). Flooding based on selective four neighbours suffers from few selection of hoping or forwarding nodes when the node density is low and does not meet the selection criteria around the source or forwarder node and may miss the chance of approaching the destination or requires large number of hops. This neighbour selection scheme uses the modified approach of the Adhoc on Demand Distance Vector (AODV) protocol to reduce the flooding of Route request (RREQ) packets, control the network overall traffic, improves link stability and the residual energy reducing the overheads by 24%.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130428914","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9924071
T. Tsuboi
This paper describes new traffic congestion analysis method challenge by social loss calculation based on actual traffic flow monitoring data in India. This actual monitoring data is collected in one of typical major city of India, where it is Ahmedabad city in Guj arat state and one of most growing cities in India. Traffic congestion is always big issues in developing counties like in India. Therefore, it is important to understand how much this traffic congestion generates impact to their life. There are several ways to measure social impact such as waste of time for commute, un-necessary fuel consumption, and air pollution and so on. In this paper, it takes a social loss which is calculated as the gap between the traffic demand and the capacity which means infrastructure supply. This social loss has potential for impact parameter of traffic congestion. In order to calculate the social loss, it is necessary to have clear traffic parameter about the traffic demand data, traffic supply data, and infrastructure cost for traffic improvement. It is not easy to get those data especially tin developing countries. This paper shows how to get traffic parameter from the actual monitoring data and how to calculate traffic social loss from this data. And as summary, this calculated social loss is valid especially for comparing the traffic congestion condition in each location. This new traffic congestion analysis by social loss is authorized by the traffic flow theory but it is the first time to bring new method for the developing countries where there are so many unknown traffic flow parameter
{"title":"Social Loss Analysis Approach of Traffic Congestion from Traffic Flow Monitoring in India","authors":"T. Tsuboi","doi":"10.1109/CSI54720.2022.9924071","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924071","url":null,"abstract":"This paper describes new traffic congestion analysis method challenge by social loss calculation based on actual traffic flow monitoring data in India. This actual monitoring data is collected in one of typical major city of India, where it is Ahmedabad city in Guj arat state and one of most growing cities in India. Traffic congestion is always big issues in developing counties like in India. Therefore, it is important to understand how much this traffic congestion generates impact to their life. There are several ways to measure social impact such as waste of time for commute, un-necessary fuel consumption, and air pollution and so on. In this paper, it takes a social loss which is calculated as the gap between the traffic demand and the capacity which means infrastructure supply. This social loss has potential for impact parameter of traffic congestion. In order to calculate the social loss, it is necessary to have clear traffic parameter about the traffic demand data, traffic supply data, and infrastructure cost for traffic improvement. It is not easy to get those data especially tin developing countries. This paper shows how to get traffic parameter from the actual monitoring data and how to calculate traffic social loss from this data. And as summary, this calculated social loss is valid especially for comparing the traffic congestion condition in each location. This new traffic congestion analysis by social loss is authorized by the traffic flow theory but it is the first time to bring new method for the developing countries where there are so many unknown traffic flow parameter","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129615143","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9924117
Anjana V Nair, Jayasree Narayanan
An inventory marketplace or a stock market is a platform for buying and selling a company's shares and derivatives at an agreed price. The goal of Stock Market Prediction is to forecast the fate fee of a company's economic shares. The latest development in market analysis technology is the use of machine earning to determine the values of current exchange indexes based on their prior values. The project entails determining the future prices of the stock markets by selecting the data from the available dataset and then determining the future pricing based on the user's desired duration of selection of months. It's built with the Prophet API and deployed with the Streamlit framework locally. Whose performance is then assessed using metrics such as Mean Squared Error and Root Mean Squared Error
库存市场或股票市场是一个以约定价格买卖公司股票和衍生品的平台。股票市场预测的目标是预测公司经济股份的命运费。市场分析技术的最新发展是利用机器学习来确定当前交易指数的价值。该项目需要通过从可用数据集中选择数据来确定股票市场的未来价格，然后根据用户期望的选择月份的持续时间来确定未来的定价。它是使用Prophet API构建的，并在本地部署了Streamlit框架。然后使用均方误差(Mean Squared Error)和均方根误差(Root Mean Squared Error)等指标评估谁的绩效
{"title":"Indian Stock Market Forecasting using Prophet Model","authors":"Anjana V Nair, Jayasree Narayanan","doi":"10.1109/CSI54720.2022.9924117","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924117","url":null,"abstract":"An inventory marketplace or a stock market is a platform for buying and selling a company's shares and derivatives at an agreed price. The goal of Stock Market Prediction is to forecast the fate fee of a company's economic shares. The latest development in market analysis technology is the use of machine earning to determine the values of current exchange indexes based on their prior values. The project entails determining the future prices of the stock markets by selecting the data from the available dataset and then determining the future pricing based on the user's desired duration of selection of months. It's built with the Prophet API and deployed with the Streamlit framework locally. Whose performance is then assessed using metrics such as Mean Squared Error and Root Mean Squared Error","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123860458","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9924012
Geetu Dhawan, A. P. Mazumdar, Y. Meena
The Internet of Things (IoT) envisions millions of devices communicating with one another. Its ubiquitous nature leads to an early dissipation of device energy, making it a hot topic for IoT research. There has been an increased interest in content-centric IoT as it facilitates critical features to constrained devices. In IoT, interest and data packet transmissions consume the majority of node energy. As the existing IoT schemes focus primarily on the node's sleep cycle, the number of transmissions, clustering, and energy harvesting, they fail to account for the size and popularity of incoming content, which increase the volume of data exchanged. Therefore, an energy-efficient strategy that considers all the parameters of constrained devices to reduce power consumption is of great importance. In this article, we proposed an energy-aware caching scheme (EAC) that caches data based on the size of incoming content while simultaneously taking into account the content's popularity and freshness. Energy-constrained IoT devices can benefit from the proposed EAC scheme, which takes into account the trade-off between content size and popularity. The experimental results show that the proposed scheme has a higher caching hit rate than the Centrally Controlled Caching (CCC) and No-Cache schemes.
{"title":"EAC: Energy-Aware Caching Scheme for Internet of Things using ICN","authors":"Geetu Dhawan, A. P. Mazumdar, Y. Meena","doi":"10.1109/CSI54720.2022.9924012","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924012","url":null,"abstract":"The Internet of Things (IoT) envisions millions of devices communicating with one another. Its ubiquitous nature leads to an early dissipation of device energy, making it a hot topic for IoT research. There has been an increased interest in content-centric IoT as it facilitates critical features to constrained devices. In IoT, interest and data packet transmissions consume the majority of node energy. As the existing IoT schemes focus primarily on the node's sleep cycle, the number of transmissions, clustering, and energy harvesting, they fail to account for the size and popularity of incoming content, which increase the volume of data exchanged. Therefore, an energy-efficient strategy that considers all the parameters of constrained devices to reduce power consumption is of great importance. In this article, we proposed an energy-aware caching scheme (EAC) that caches data based on the size of incoming content while simultaneously taking into account the content's popularity and freshness. Energy-constrained IoT devices can benefit from the proposed EAC scheme, which takes into account the trade-off between content size and popularity. The experimental results show that the proposed scheme has a higher caching hit rate than the Centrally Controlled Caching (CCC) and No-Cache schemes.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125525750","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9924079
N. Kasthuri, R. Ramyea, D. Jeffrin, N. K. Chitrasena, K. Divveshwari
In India, over 3 to 4 % of people are affected by chronic proliferative disease known as Psoriasis. The patchy rashes, small scaling spots, cracked skin, itching, burning or soreness are the common signs and symptoms of psoriasis. These signs and symptoms vary depending on the type of psoriasis. The various types of psoriasis are plague psoriasis, nail psoriasis, Guttate psoriasis, inverse, pustular and erythrodermic psoriasis. These psoriasis affects the skin and, in some cases, the fungal infections will trigger the disease. To evaluate psoriasis severity, various methods are used to monitor the therapeutic response. In this paper, Principal Component Analysis (PCA) and rigid transformations are used for the automatic segmentation of psoriasis. Convolutional Neural Network (CNN) which comprises of convolutional layer, ReLU activation layer, max pooling layer and fully connected feed forward network are used for the classification of skin images. The feature map is extracted from the input images by convolution operation. These feature maps are obtained to train the neural network model to classify the images. The performance metric of the model is calculated after training the model with input images and the model performance varies depending on the type of images. The accuracy, specificity, sensitivity, F1 score are determined to find the best model for evaluation.
{"title":"CNN Based Automatic Segmentation of Scaling in 2-D Psoriasis Skin Images","authors":"N. Kasthuri, R. Ramyea, D. Jeffrin, N. K. Chitrasena, K. Divveshwari","doi":"10.1109/CSI54720.2022.9924079","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924079","url":null,"abstract":"In India, over 3 to 4 % of people are affected by chronic proliferative disease known as Psoriasis. The patchy rashes, small scaling spots, cracked skin, itching, burning or soreness are the common signs and symptoms of psoriasis. These signs and symptoms vary depending on the type of psoriasis. The various types of psoriasis are plague psoriasis, nail psoriasis, Guttate psoriasis, inverse, pustular and erythrodermic psoriasis. These psoriasis affects the skin and, in some cases, the fungal infections will trigger the disease. To evaluate psoriasis severity, various methods are used to monitor the therapeutic response. In this paper, Principal Component Analysis (PCA) and rigid transformations are used for the automatic segmentation of psoriasis. Convolutional Neural Network (CNN) which comprises of convolutional layer, ReLU activation layer, max pooling layer and fully connected feed forward network are used for the classification of skin images. The feature map is extracted from the input images by convolution operation. These feature maps are obtained to train the neural network model to classify the images. The performance metric of the model is calculated after training the model with input images and the model performance varies depending on the type of images. The accuracy, specificity, sensitivity, F1 score are determined to find the best model for evaluation.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127198981","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9924088
P. Priya, Sushma Kamlu
Recent advances in sensor technologies, actuators, and power storage have opened the door to the development of crewless flying vehicles. The information in this paper is on quadrotor modeling and control. As a result, in this study, a mathematical analysis of the Quad-rotor Unmanned Aerial Vehicle (UAV) is first developed, which is on the basis of the Euler-Newton approach of equating motion and moment forces. Second, a Robust Proportional Integral (PI) system is used to robust the quad-rotor UAV's height and attitude. The PI coefficients are optimized using the Improved Genetic Algorithm (Improve-GA) technique to generate an acceptable outcome for the system. After running several simulations, it was determined that the PI controller is able to monitor the appropriate reference values. According to experimental data, the suggested improved GA-PI controller quickly and effectively stabilized a Quad-rotor UAV, and its response time and settling time are reasonable for attitude stabilization control applications.
{"title":"Improved GA-PI Technique for Non-Linear Dynamic Modelling of a UAV","authors":"P. Priya, Sushma Kamlu","doi":"10.1109/CSI54720.2022.9924088","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924088","url":null,"abstract":"Recent advances in sensor technologies, actuators, and power storage have opened the door to the development of crewless flying vehicles. The information in this paper is on quadrotor modeling and control. As a result, in this study, a mathematical analysis of the Quad-rotor Unmanned Aerial Vehicle (UAV) is first developed, which is on the basis of the Euler-Newton approach of equating motion and moment forces. Second, a Robust Proportional Integral (PI) system is used to robust the quad-rotor UAV's height and attitude. The PI coefficients are optimized using the Improved Genetic Algorithm (Improve-GA) technique to generate an acceptable outcome for the system. After running several simulations, it was determined that the PI controller is able to monitor the appropriate reference values. According to experimental data, the suggested improved GA-PI controller quickly and effectively stabilized a Quad-rotor UAV, and its response time and settling time are reasonable for attitude stabilization control applications.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"74 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127643279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9924066
Indranil Sarkar, Sanjay Kumar
In the era of Internet of Things (loT), data offloading become a promising and crucial strategy to improve the overall system performance and also to provide quality-of-service (QOS). In this context, recently fog computing has gained a lot of interests from the industry as well as academia. In this paper, we propose a delay-aware task offloading strategy in mobile fog-based network. We consider several moving vehicles in a one-way road out of which some vehicles act as client vehicles and some of them act as mobile fog nodes. Individual fog nodes allocate its available resources to the the requesting client vehicles in its proximity. However, because of the dynamic nature of the vehicular environment, it is difficult to develop a scheme that can decide how to allocate the computing resources to the local on-board CPU or to the neighbouring fog nodes. In this regards, the paper propose a deep reinforcement learning based intelligent task offloading for vehicles in motion (ITOVM) policy, considering the vehicle mobility and communication bandwidth constraints, to minimize the overall latency of the network. The proposed IOTVM policy is formulated as the Markov decision process (MDP) which is solved by the concept of deep Q network (DQN). Finally, extensive simulation results demonstrate the efficacy and performance enhancement of the proposed approach compared to several baseline algorithms.
在物联网(loT)时代，数据卸载成为提高系统整体性能和提供服务质量(QOS)的重要策略。在这种背景下，最近雾计算在工业界和学术界引起了很大的兴趣。本文提出了一种基于移动雾的网络延迟感知任务卸载策略。我们考虑几辆在单行道上移动的车辆，其中一些车辆充当客户端车辆，其中一些充当移动雾节点。单个雾节点将其可用资源分配给其附近的请求客户端车辆。然而，由于车辆环境的动态性，很难制定一种方案来决定如何将计算资源分配给本地车载CPU或相邻雾节点。为此，本文提出了一种基于深度强化学习的ITOVM (intelligent task offloading for vehicles In motion)策略，考虑到车辆移动性和通信带宽的限制，以最小化网络的整体延迟。所提出的IOTVM策略被表述为马尔可夫决策过程(MDP)，并通过深度Q网络(DQN)的概念进行求解。最后，大量的仿真结果表明，与几种基准算法相比，该方法的有效性和性能增强。
{"title":"Delay-aware Intelligent Task Offloading Strategy in Vehicular Fog Computing","authors":"Indranil Sarkar, Sanjay Kumar","doi":"10.1109/CSI54720.2022.9924066","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924066","url":null,"abstract":"In the era of Internet of Things (loT), data offloading become a promising and crucial strategy to improve the overall system performance and also to provide quality-of-service (QOS). In this context, recently fog computing has gained a lot of interests from the industry as well as academia. In this paper, we propose a delay-aware task offloading strategy in mobile fog-based network. We consider several moving vehicles in a one-way road out of which some vehicles act as client vehicles and some of them act as mobile fog nodes. Individual fog nodes allocate its available resources to the the requesting client vehicles in its proximity. However, because of the dynamic nature of the vehicular environment, it is difficult to develop a scheme that can decide how to allocate the computing resources to the local on-board CPU or to the neighbouring fog nodes. In this regards, the paper propose a deep reinforcement learning based intelligent task offloading for vehicles in motion (ITOVM) policy, considering the vehicle mobility and communication bandwidth constraints, to minimize the overall latency of the network. The proposed IOTVM policy is formulated as the Markov decision process (MDP) which is solved by the concept of deep Q network (DQN). Finally, extensive simulation results demonstrate the efficacy and performance enhancement of the proposed approach compared to several baseline algorithms.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130125238","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9923970
S. T. S., P. Sreeja
Millions of individuals throughout the world consider several social networking. The Network sites/apps, such as WhatsApp, Instagram, Twitter, and Facebook, are essential information sources due to their vast user bases. Social networking platforms provide hands-on connectivity with others and ease of use to the down-market that radio and television broadcasters are unable to provide - multi-way communication. By modifying the original content or a full parody presented as fact, negative influencers can take advantage of the freedom provided by social networks and wilfully spread any misinformation. The spread of false information can happen in the blink of an eye, potentially deceiving the general public and spreading to other communities. Despite this awareness, social media platforms continue to spread dubious material. Even though many of them are still passed off as fact, hoaxes continue to be a major problem. To get around this problem, this study proposed a potent technique called BERT-LSTM to identify false information on Facebook. The accuracy of the suggested BERT-LSTM approach is 0.828.
{"title":"BERT-LSTM for Fake News Detection on Facebook Using SVD","authors":"S. T. S., P. Sreeja","doi":"10.1109/CSI54720.2022.9923970","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9923970","url":null,"abstract":"Millions of individuals throughout the world consider several social networking. The Network sites/apps, such as WhatsApp, Instagram, Twitter, and Facebook, are essential information sources due to their vast user bases. Social networking platforms provide hands-on connectivity with others and ease of use to the down-market that radio and television broadcasters are unable to provide - multi-way communication. By modifying the original content or a full parody presented as fact, negative influencers can take advantage of the freedom provided by social networks and wilfully spread any misinformation. The spread of false information can happen in the blink of an eye, potentially deceiving the general public and spreading to other communities. Despite this awareness, social media platforms continue to spread dubious material. Even though many of them are still passed off as fact, hoaxes continue to be a major problem. To get around this problem, this study proposed a potent technique called BERT-LSTM to identify false information on Facebook. The accuracy of the suggested BERT-LSTM approach is 0.828.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"72 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124240255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9924044
Chandana Udupa, Anusha Upadhyaya, Basanagoud S. Patil, S. Seeri, Prakashgoud Patil, P. Hiremath
Text detection and its script identification in a natural scene image/video has attracted the attention of many researchers over the recent years due to its application in the de-sign of computer vision devices for usage by the visually impaired people, global tourists travelling in unfamiliar tourist places, etc. to facilitate them to understand the textual information displayed on sign boards, bill boards, public notice boards, etc., the objective of the proposed method is detection and localization of multilingual text in a natural scene video image and its corresponding script identification. The texts in three languages, namely, English, Hindi and Kannada, are considered. In the proposed method, CNN based YOLOv5 is used for text detection and localization in real-time videos of natural scene and it is also trained for script identification. The YOLOv5 performance is found to yield an accuracy higher than otherobject detection algorithms. The proposed model is trained witha custom dataset containing video images of natural scenes and istested for different scenarios like texts in different backgrounds, fonts, orientations, resolutions, and disturbances in the images. The experimental results demonstrate the effectiveness and robustness of the proposed method. The performance comparison is done with other methods in the literature.
{"title":"Text Localization and Script Identification in Natural Scene Images and Videos","authors":"Chandana Udupa, Anusha Upadhyaya, Basanagoud S. Patil, S. Seeri, Prakashgoud Patil, P. Hiremath","doi":"10.1109/CSI54720.2022.9924044","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924044","url":null,"abstract":"Text detection and its script identification in a natural scene image/video has attracted the attention of many researchers over the recent years due to its application in the de-sign of computer vision devices for usage by the visually impaired people, global tourists travelling in unfamiliar tourist places, etc. to facilitate them to understand the textual information displayed on sign boards, bill boards, public notice boards, etc., the objective of the proposed method is detection and localization of multilingual text in a natural scene video image and its corresponding script identification. The texts in three languages, namely, English, Hindi and Kannada, are considered. In the proposed method, CNN based YOLOv5 is used for text detection and localization in real-time videos of natural scene and it is also trained for script identification. The YOLOv5 performance is found to yield an accuracy higher than otherobject detection algorithms. The proposed model is trained witha custom dataset containing video images of natural scenes and istested for different scenarios like texts in different backgrounds, fonts, orientations, resolutions, and disturbances in the images. The experimental results demonstrate the effectiveness and robustness of the proposed method. The performance comparison is done with other methods in the literature.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"250 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117171914","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：