With the vigorous development of computing power, Convolutional Neural Network (CNN) is developing rapidly, and new CNN structures with more layers and better performance continue to appear. Field Programmable Gate Array(FPGA) has gradually become the best choice for people to deploy and accelerate CNNs as a current research hotspot. This paper has studied the hardware acceleration method of FPGA to implement and simulate the Softmax layer of Alexnet on Vivado 2018.1. Combined with the features of FPGA, the Cordic algorithm is used to implement basic operations such as division and exponential functions, instead of consuming floating-point arithmetic resources. The paper proposes a method to shrink the convergence domain and analyzes the errors generated by the different digits of data after quantization and fixed-point inputs. The relative error of the Softmax layer exponential function is controlled below 0.0146% by reducing the bit width which satisfied the design requirements and saved resources. This method can complete the calculation and classification of the Softmax layer in 66.5 cycles without processing the layer data at fixed points, which greatly improves the calculation speed of the Softmax layer.
{"title":"Cordic-based Softmax Acceleration Method of Convolution Neural Network on FPGA","authors":"Yongxiang Cao, Wan'ang Xiao, Jingdun Jia, Dehua Wu, Weixin Zhou","doi":"10.1109/ICAIIS49377.2020.9194894","DOIUrl":"https://doi.org/10.1109/ICAIIS49377.2020.9194894","url":null,"abstract":"With the vigorous development of computing power, Convolutional Neural Network (CNN) is developing rapidly, and new CNN structures with more layers and better performance continue to appear. Field Programmable Gate Array(FPGA) has gradually become the best choice for people to deploy and accelerate CNNs as a current research hotspot. This paper has studied the hardware acceleration method of FPGA to implement and simulate the Softmax layer of Alexnet on Vivado 2018.1. Combined with the features of FPGA, the Cordic algorithm is used to implement basic operations such as division and exponential functions, instead of consuming floating-point arithmetic resources. The paper proposes a method to shrink the convergence domain and analyzes the errors generated by the different digits of data after quantization and fixed-point inputs. The relative error of the Softmax layer exponential function is controlled below 0.0146% by reducing the bit width which satisfied the design requirements and saved resources. This method can complete the calculation and classification of the Softmax layer in 66.5 cycles without processing the layer data at fixed points, which greatly improves the calculation speed of the Softmax layer.","PeriodicalId":416002,"journal":{"name":"2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124058692","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 : 2020-03-01DOI: 10.1109/ICAIIS49377.2020.9194911
P. Duan, Jiahao Pan, Wenbi Rao
Scene text detection and scene text recognition are important components of scene text recognition system. Scene text detection, the initial stage of scene text recognition, aims to find out text area in the picture. Recently the target detection method Mask R-CNN has been employed scene text detection and achieved good performance. In this paper, we set forth a model, MaskS R-CNN text detector, based on Mask R-CNN, which attempts to detect scene text. In this model, a network block of Mask Scoring R-CNN is introduced to learn the high quality of the predicted instance mask scores. The mask scoring mechanism correct the inconformity between mask quality and mask score, at the same time improves instance segmentation performance by attaching great importance to more accurate mask predictions. The method put forward in this paper can achieve multi-directional and multi-language natural scene text detection. Compared with some existing traditional location methods based on edge, color and texture and some location methods based on deep learning, it is a relatively innovative method.
{"title":"MaskS R-CNN Text Detector","authors":"P. Duan, Jiahao Pan, Wenbi Rao","doi":"10.1109/ICAIIS49377.2020.9194911","DOIUrl":"https://doi.org/10.1109/ICAIIS49377.2020.9194911","url":null,"abstract":"Scene text detection and scene text recognition are important components of scene text recognition system. Scene text detection, the initial stage of scene text recognition, aims to find out text area in the picture. Recently the target detection method Mask R-CNN has been employed scene text detection and achieved good performance. In this paper, we set forth a model, MaskS R-CNN text detector, based on Mask R-CNN, which attempts to detect scene text. In this model, a network block of Mask Scoring R-CNN is introduced to learn the high quality of the predicted instance mask scores. The mask scoring mechanism correct the inconformity between mask quality and mask score, at the same time improves instance segmentation performance by attaching great importance to more accurate mask predictions. The method put forward in this paper can achieve multi-directional and multi-language natural scene text detection. Compared with some existing traditional location methods based on edge, color and texture and some location methods based on deep learning, it is a relatively innovative method.","PeriodicalId":416002,"journal":{"name":"2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130339114","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 : 2020-03-01DOI: 10.1109/ICAIIS49377.2020.9194919
Xiaoping Yu, Lin Zhu, Lanxu Jia
The system collects and detects hand information of relevant persons in a specific area, and transmits the data to the server for calculation and processing. In response to the abnormal state of the hand, the result picture is fed back to the staff in real time through the mobile app. This paper proposes a method for detecting and recognizing abnormal hand states based on the improved yolov3 algorithm. The system collects real-time pictures of the hand through the camera to determine whether the hand is carrying ring, bandages, and whether there are bleeding points. After optimizing the network and preprocessing the data, the algorithm accuracy can reach 99.7%. In addition, the simplified processing of the model can reduce the burden on the hardware system.
{"title":"Detection and recognition of hand abnormal state based on deep learning algorithm","authors":"Xiaoping Yu, Lin Zhu, Lanxu Jia","doi":"10.1109/ICAIIS49377.2020.9194919","DOIUrl":"https://doi.org/10.1109/ICAIIS49377.2020.9194919","url":null,"abstract":"The system collects and detects hand information of relevant persons in a specific area, and transmits the data to the server for calculation and processing. In response to the abnormal state of the hand, the result picture is fed back to the staff in real time through the mobile app. This paper proposes a method for detecting and recognizing abnormal hand states based on the improved yolov3 algorithm. The system collects real-time pictures of the hand through the camera to determine whether the hand is carrying ring, bandages, and whether there are bleeding points. After optimizing the network and preprocessing the data, the algorithm accuracy can reach 99.7%. In addition, the simplified processing of the model can reduce the burden on the hardware system.","PeriodicalId":416002,"journal":{"name":"2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS)","volume":"189 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128934783","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 : 2020-03-01DOI: 10.1109/ICAIIS49377.2020.9194809
Cossi Blaise Avoussoukpo, Chunxiang Xu, Marius Tchenagnon
The concept of the Internet of things is a paradigm that is supposed to integrate several complex technologies. Opportunistic Networks due to its nature, do share common characteristics with the Internet of Things. Thus, the study of opportunistic Networks could greatly benefit the Internet of Things area of research. The Opportunistic Networks aim to use the altruism of mobile or immobile things to achieve specific tasks. However, among other challenges to solve for the Opportunistic Networks to work appropriately is the lack of a flexible, low-cost, and polyvalent communication system. Thankfully, the Polyvalent(Multipurpose) Wireless Communication System (PWCS) appears as a serious candidate to solve that challenge. This paper aims to show the correlation between an existing but unexplored technology; the PWCS, and Opportunistic Networks providing four main contributions. First, this paper clarifies the concept of opportunistic networks. Second, it points out the key differences between OppNets and some communications models that emerged from Mobile Ad hoc Networks research. Third, it presents the PWCS. Finally, it provides the correlation between OppNets and the PWCS and also discusses the limitations of the PWCS. The PWCS, if given unbiased attention, may add value to the communication field.
{"title":"Polyvalent Wireless Communication System (PWCS); A Potentially Useful Technology for Opportunistic Networks","authors":"Cossi Blaise Avoussoukpo, Chunxiang Xu, Marius Tchenagnon","doi":"10.1109/ICAIIS49377.2020.9194809","DOIUrl":"https://doi.org/10.1109/ICAIIS49377.2020.9194809","url":null,"abstract":"The concept of the Internet of things is a paradigm that is supposed to integrate several complex technologies. Opportunistic Networks due to its nature, do share common characteristics with the Internet of Things. Thus, the study of opportunistic Networks could greatly benefit the Internet of Things area of research. The Opportunistic Networks aim to use the altruism of mobile or immobile things to achieve specific tasks. However, among other challenges to solve for the Opportunistic Networks to work appropriately is the lack of a flexible, low-cost, and polyvalent communication system. Thankfully, the Polyvalent(Multipurpose) Wireless Communication System (PWCS) appears as a serious candidate to solve that challenge. This paper aims to show the correlation between an existing but unexplored technology; the PWCS, and Opportunistic Networks providing four main contributions. First, this paper clarifies the concept of opportunistic networks. Second, it points out the key differences between OppNets and some communications models that emerged from Mobile Ad hoc Networks research. Third, it presents the PWCS. Finally, it provides the correlation between OppNets and the PWCS and also discusses the limitations of the PWCS. The PWCS, if given unbiased attention, may add value to the communication field.","PeriodicalId":416002,"journal":{"name":"2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127604702","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 : 2020-03-01DOI: 10.1109/ICAIIS49377.2020.9194867
Yueqing Li, Shihui Zheng
As a further development of the side channel technique, the security of crypto-chips has received extensive attention from researchers. A mask scheme based on a random number generator is usually used to protect crypto-chips, but it will result in extra memory and time consumption. How to balance the security of crypto-chips and the memory consumption has always been the research focus. In this paper, we use a random number generator to design a random loop mask scheme (RLM) for GIFT algorithm. The binary value of a random number is used to set the position and value of masks. In addition, the rotating S-box of the RSM scheme is combined with the small generation mask of a fixed mask scheme to make the RLM more secure. The experimental results show that the RLM scheme protects the intermediate values from being leaked. Compared to other common schemes, our RLM scheme could resist not only correlation power analysis (CPA) attacks but also higher-order correlation power analysis (HO-CP A) attacks. In addition, the RAM consumption was 49% less than that of the global mask scheme.
{"title":"RLM: a new mask countermeasure to resist HO-CPA for GIFT","authors":"Yueqing Li, Shihui Zheng","doi":"10.1109/ICAIIS49377.2020.9194867","DOIUrl":"https://doi.org/10.1109/ICAIIS49377.2020.9194867","url":null,"abstract":"As a further development of the side channel technique, the security of crypto-chips has received extensive attention from researchers. A mask scheme based on a random number generator is usually used to protect crypto-chips, but it will result in extra memory and time consumption. How to balance the security of crypto-chips and the memory consumption has always been the research focus. In this paper, we use a random number generator to design a random loop mask scheme (RLM) for GIFT algorithm. The binary value of a random number is used to set the position and value of masks. In addition, the rotating S-box of the RSM scheme is combined with the small generation mask of a fixed mask scheme to make the RLM more secure. The experimental results show that the RLM scheme protects the intermediate values from being leaked. Compared to other common schemes, our RLM scheme could resist not only correlation power analysis (CPA) attacks but also higher-order correlation power analysis (HO-CP A) attacks. In addition, the RAM consumption was 49% less than that of the global mask scheme.","PeriodicalId":416002,"journal":{"name":"2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS)","volume":"85 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127850504","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 : 2020-03-01DOI: 10.1109/ICAIIS49377.2020.9194831
N. Echoda, S. Z. Farooq, Hong Xuebao, J. Chukwuma, Dongkai Yang
Multipath being a stochastic GNSS error is hard to eliminate. For two receivers operating in differential GNSS mode with a short baseline (below 10 km in this case), it is possible to mitigate geographically common errors to both receivers but not multipath. Real Time Kinematic (RTK) uses carrier-phase observations to form differences in real time. The single frequency (SF) code observations for receivers functioning in RTK are filtered through code-carrier combination of observations. The Hatch filter algorithm is often utilized in SF-GNSS based applications to detect cycle slips and also to mitigate multipath. However, choosing an effective smoothing window is important to ensure that the Hatch filtered output matches closely to that of the input raw signal. Dynamic Time Warping (DTW) was used to get appropriate smoothing constant value for the Hatch filter by measuring the squared Euclidean distances of corresponding points of the two signals. Using dynamic data collected for 906 epochs from two receivers configured in RTK mode, a bespoke SF-RTK software was used to compare the positioning results of the Hatch filtered output to that of the widely used RTKLIB software. The results showed that the standard deviation of the differences of the ECEF coordinates between the two software are within centimetre levels.
多路径误差是GNSS的随机误差，难以消除。对于两个以差分GNSS模式运行的接收器，其基线较短(在这种情况下低于10公里)，可以减轻两个接收器在地理上的常见错误，但不能减轻多路径。实时运动学(Real Time Kinematic, RTK)利用载波相位观测来实时形成差分。在RTK中运行的接收机的单频(SF)代码观测值通过观测值的码载波组合进行过滤。哈奇滤波算法通常用于基于SF-GNSS的应用中，以检测周跳和减轻多径。然而，选择一个有效的平滑窗口是重要的，以确保Hatch滤波输出与输入原始信号紧密匹配。通过测量两个信号对应点的欧氏距离的平方，采用动态时间扭曲(Dynamic Time Warping, DTW)得到合适的Hatch滤波器平滑常数值。利用配置为RTK模式的两个接收机906次采集的动态数据，使用定制的SF-RTK软件将Hatch滤波输出的定位结果与广泛使用的RTKLIB软件的定位结果进行比较。结果表明，两种软件ECEF坐标差值的标准差均在厘米级以内。
{"title":"Multipath Mitigation Analysis using Hatch Filter and DTW in Single Frequency RTK","authors":"N. Echoda, S. Z. Farooq, Hong Xuebao, J. Chukwuma, Dongkai Yang","doi":"10.1109/ICAIIS49377.2020.9194831","DOIUrl":"https://doi.org/10.1109/ICAIIS49377.2020.9194831","url":null,"abstract":"Multipath being a stochastic GNSS error is hard to eliminate. For two receivers operating in differential GNSS mode with a short baseline (below 10 km in this case), it is possible to mitigate geographically common errors to both receivers but not multipath. Real Time Kinematic (RTK) uses carrier-phase observations to form differences in real time. The single frequency (SF) code observations for receivers functioning in RTK are filtered through code-carrier combination of observations. The Hatch filter algorithm is often utilized in SF-GNSS based applications to detect cycle slips and also to mitigate multipath. However, choosing an effective smoothing window is important to ensure that the Hatch filtered output matches closely to that of the input raw signal. Dynamic Time Warping (DTW) was used to get appropriate smoothing constant value for the Hatch filter by measuring the squared Euclidean distances of corresponding points of the two signals. Using dynamic data collected for 906 epochs from two receivers configured in RTK mode, a bespoke SF-RTK software was used to compare the positioning results of the Hatch filtered output to that of the widely used RTKLIB software. The results showed that the standard deviation of the differences of the ECEF coordinates between the two software are within centimetre levels.","PeriodicalId":416002,"journal":{"name":"2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122719627","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}
Because the electric vehicle load is a complex space-time load with comprehensive influence factors. Therefore, this paper considers the temporal and spatial characteristics of the electric vehicle load and the time sequence characteristics of the distributed generation. It establishes optimization model of distribution network containing the plug-in electric vehicle and distributed generation. To optimize the annual total cost of distribution network, it uses Monte Carlo simulation to get the temporal and spatial characteristics of electric vehicle load. Based on the coordinated control strategy, the model is solved by a hybrid particle swarm optimization algorithm with mutation and crossover operation.
{"title":"Research on the optimal coordinated control strategy of ‘source-grid-load-storage’ including electric vehicle and distributed power supply","authors":"Guang-bin Huang, Hao Ma, Xi Chen, Binrong Zhang, Chang Liu, Jing Zhang","doi":"10.1109/ICAIIS49377.2020.9194707","DOIUrl":"https://doi.org/10.1109/ICAIIS49377.2020.9194707","url":null,"abstract":"Because the electric vehicle load is a complex space-time load with comprehensive influence factors. Therefore, this paper considers the temporal and spatial characteristics of the electric vehicle load and the time sequence characteristics of the distributed generation. It establishes optimization model of distribution network containing the plug-in electric vehicle and distributed generation. To optimize the annual total cost of distribution network, it uses Monte Carlo simulation to get the temporal and spatial characteristics of electric vehicle load. Based on the coordinated control strategy, the model is solved by a hybrid particle swarm optimization algorithm with mutation and crossover operation.","PeriodicalId":416002,"journal":{"name":"2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123091696","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 : 2020-03-01DOI: 10.1109/ICAIIS49377.2020.9194881
Yong Wang, Ling Li, Xin Yang, Xinxin Wang, Hui Liu
Camouflaged object detection is a hard assignment due to their textures are similar to the background. The main intention of this paper is probe into a problem about the camouflaged object detection, that is, detecting its camouflaged object for a given image. This problem has not been well studied in spite of a large area of potential applications such as camouflage military targets detection and wildlife protection. To address this problem, a camouflage object detection method based on deep learning is proposed. The suggested method can detect camouflaged object which can extract deep features automatically. It can also provide detection probability which reflect camouflage efficiency. Experimental results show that the deep learning measure can effectively detect different scene, representing the camouflage level of low, medium and high respectively.
{"title":"A Camouflaged Object Detection Model Based on Deep Learning","authors":"Yong Wang, Ling Li, Xin Yang, Xinxin Wang, Hui Liu","doi":"10.1109/ICAIIS49377.2020.9194881","DOIUrl":"https://doi.org/10.1109/ICAIIS49377.2020.9194881","url":null,"abstract":"Camouflaged object detection is a hard assignment due to their textures are similar to the background. The main intention of this paper is probe into a problem about the camouflaged object detection, that is, detecting its camouflaged object for a given image. This problem has not been well studied in spite of a large area of potential applications such as camouflage military targets detection and wildlife protection. To address this problem, a camouflage object detection method based on deep learning is proposed. The suggested method can detect camouflaged object which can extract deep features automatically. It can also provide detection probability which reflect camouflage efficiency. Experimental results show that the deep learning measure can effectively detect different scene, representing the camouflage level of low, medium and high respectively.","PeriodicalId":416002,"journal":{"name":"2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS)","volume":"322 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115771783","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 : 2020-03-01DOI: 10.1109/ICAIIS49377.2020.9194818
Sui He, Ding Lei, Wang Shuang, Chunbo Liu, Zhaojun Gu
In order to cope with the network attack of industrial control system, this paper proposes a quantifiable attack-defense tree model. In order to reduce the influence of subjective factors on weight calculation and the probability of attack events, the Fuzzy Analytic Hierarchy Process and the Attack-Defense Tree model are combined. First, the model provides a variety of security attributes for attack and defense leaf nodes. Secondly, combining the characteristics of leaf nodes, a fuzzy consistency matrix is constructed to calculate the security attribute weight of leaf nodes, and the probability of attack and defense leaf nodes. Then, the influence of defense node on attack behavior is analyzed. Finally, the network risk assessment of typical airport oil supply automatic control system has been undertaken as a case study using this attack-defense tree model. The result shows that this model can truly reflect the impact of defense measures on the attack behavior, and provide a reference for the network security scheme.
{"title":"Network Security Analysis of Industrial Control System Based on Attack-Defense Tree","authors":"Sui He, Ding Lei, Wang Shuang, Chunbo Liu, Zhaojun Gu","doi":"10.1109/ICAIIS49377.2020.9194818","DOIUrl":"https://doi.org/10.1109/ICAIIS49377.2020.9194818","url":null,"abstract":"In order to cope with the network attack of industrial control system, this paper proposes a quantifiable attack-defense tree model. In order to reduce the influence of subjective factors on weight calculation and the probability of attack events, the Fuzzy Analytic Hierarchy Process and the Attack-Defense Tree model are combined. First, the model provides a variety of security attributes for attack and defense leaf nodes. Secondly, combining the characteristics of leaf nodes, a fuzzy consistency matrix is constructed to calculate the security attribute weight of leaf nodes, and the probability of attack and defense leaf nodes. Then, the influence of defense node on attack behavior is analyzed. Finally, the network risk assessment of typical airport oil supply automatic control system has been undertaken as a case study using this attack-defense tree model. The result shows that this model can truly reflect the impact of defense measures on the attack behavior, and provide a reference for the network security scheme.","PeriodicalId":416002,"journal":{"name":"2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131193463","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 : 2020-03-01DOI: 10.1109/ICAIIS49377.2020.9194918
Xuejun Wang, Zhiguo Zhou, Yun Li
This paper proposes an improved ASMS algorithm for mean shift scale problem, which integrates the ASMS algorithm into the Ross system and transplants it to the Jetsontx2 platform. In this method, the outline rectangle frame of the object to be tracked is drawn on the picture sent back by UAV, and the pixel coordinates of the rectangle frame are encapsulated into Mavlink protocol and sent to Jetsontx2 platform through UDP for tracking algorithm. The experimental results show that the tracking system based on Jetsontx2 platform can overcome the problems of tracking instability, poor imaging effect and low definition.
{"title":"Design of moving target tracking system based on Jetson platform","authors":"Xuejun Wang, Zhiguo Zhou, Yun Li","doi":"10.1109/ICAIIS49377.2020.9194918","DOIUrl":"https://doi.org/10.1109/ICAIIS49377.2020.9194918","url":null,"abstract":"This paper proposes an improved ASMS algorithm for mean shift scale problem, which integrates the ASMS algorithm into the Ross system and transplants it to the Jetsontx2 platform. In this method, the outline rectangle frame of the object to be tracked is drawn on the picture sent back by UAV, and the pixel coordinates of the rectangle frame are encapsulated into Mavlink protocol and sent to Jetsontx2 platform through UDP for tracking algorithm. The experimental results show that the tracking system based on Jetsontx2 platform can overcome the problems of tracking instability, poor imaging effect and low definition.","PeriodicalId":416002,"journal":{"name":"2020 IEEE International Conference on Artificial Intelligence and Information Systems (ICAIIS)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123804976","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：