Pub Date : 2021-08-01DOI: 10.1109/ISCEIC53685.2021.00098
M. Islam, Khandaker Sajidul Islam, Abdullah Noman, Joseph Ncube, Xiaohua Chen
A virtualized multi-domain architecture supports 5G fixed-line communications. The conventional virtual network display software creates a lot of surpluses and overlays communications given a large number of fundamental network access. At the same time, numerous internet infrastructure providers are reluctant to communicate specific spatial information in consideration of economic interests. Therefore, based on the resource citation, monomeric virtual network coordination is created to optimize network resources seamlessly for every autonomous domain. Multiple connections to the virtual network single connection instruction under the premise of the VNE fiber environment and the difference between the virtual throughput can be established to decrease the cost of the implementation of the virtual network. This study found that using multiple links to single-link network connections improved the 5G network-based VNE system. Discover the advantages of multiple links to a single link network as well as static vs dynamic network configuration. First of all, it's important to consider various link resources and inter-domain connectivity capabilities for VNE fiber-based wi-fi networks. The fibers connectivity virtual network integration process is provided for efficient routing depending on the perfectly designed predator algorithm. In the second step, an algorithm with a relatively small overhead based on inter-domain maps. The simulations show that the wireless fiber-based play a fundamental role for numerous systems to multi-dominate multiple virtual networks to single domain network can embedded systems can minimize the cost of embedding the virtual network embedding of multi-domain systems and have high stability and growth.
{"title":"A Fiber Wireless Improved 5G Network-Based Virtual Networking System Focused on Equal Bandwidth","authors":"M. Islam, Khandaker Sajidul Islam, Abdullah Noman, Joseph Ncube, Xiaohua Chen","doi":"10.1109/ISCEIC53685.2021.00098","DOIUrl":"https://doi.org/10.1109/ISCEIC53685.2021.00098","url":null,"abstract":"A virtualized multi-domain architecture supports 5G fixed-line communications. The conventional virtual network display software creates a lot of surpluses and overlays communications given a large number of fundamental network access. At the same time, numerous internet infrastructure providers are reluctant to communicate specific spatial information in consideration of economic interests. Therefore, based on the resource citation, monomeric virtual network coordination is created to optimize network resources seamlessly for every autonomous domain. Multiple connections to the virtual network single connection instruction under the premise of the VNE fiber environment and the difference between the virtual throughput can be established to decrease the cost of the implementation of the virtual network. This study found that using multiple links to single-link network connections improved the 5G network-based VNE system. Discover the advantages of multiple links to a single link network as well as static vs dynamic network configuration. First of all, it's important to consider various link resources and inter-domain connectivity capabilities for VNE fiber-based wi-fi networks. The fibers connectivity virtual network integration process is provided for efficient routing depending on the perfectly designed predator algorithm. In the second step, an algorithm with a relatively small overhead based on inter-domain maps. The simulations show that the wireless fiber-based play a fundamental role for numerous systems to multi-dominate multiple virtual networks to single domain network can embedded systems can minimize the cost of embedding the virtual network embedding of multi-domain systems and have high stability and growth.","PeriodicalId":342968,"journal":{"name":"2021 2nd International Symposium on Computer Engineering and Intelligent Communications (ISCEIC)","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121709981","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 : 2021-08-01DOI: 10.1109/ISCEIC53685.2021.00014
Binyu Wang, Lijie Li
Aiming at the problem of path selection error and the prior information of channel sparsity in compressed sampling matching pursuit algorithm, an improved sparsity adaptive compressed sensing algorithm based on atomic screening is proposed. The improved algorithm uses the time-domain screening characteristics of traditional DFT channel estimation algorithm to screen out most noise atoms, then uses the atom set obtained as priori information to reduce the probability of error selection in traditional compressed sensing algorithm. In addition, the idea of step size is introduced. The number of atoms entering the candidate set in each iteration is set to twice the step size, further reducing the possibility of irrelevant atoms entering the candidate set. Finally, the number of iterations is controlled by the change of residual, which improves the adaptability of channel estimation algorithm. The simulation results show that the mean square error performance of the improved algorithm in the channel with unknown sparsity is 4dB higher than that of the original algorithm.
{"title":"Sparsity Adaptive Channel Estimation Algorithm Based on Compressed Sensing","authors":"Binyu Wang, Lijie Li","doi":"10.1109/ISCEIC53685.2021.00014","DOIUrl":"https://doi.org/10.1109/ISCEIC53685.2021.00014","url":null,"abstract":"Aiming at the problem of path selection error and the prior information of channel sparsity in compressed sampling matching pursuit algorithm, an improved sparsity adaptive compressed sensing algorithm based on atomic screening is proposed. The improved algorithm uses the time-domain screening characteristics of traditional DFT channel estimation algorithm to screen out most noise atoms, then uses the atom set obtained as priori information to reduce the probability of error selection in traditional compressed sensing algorithm. In addition, the idea of step size is introduced. The number of atoms entering the candidate set in each iteration is set to twice the step size, further reducing the possibility of irrelevant atoms entering the candidate set. Finally, the number of iterations is controlled by the change of residual, which improves the adaptability of channel estimation algorithm. The simulation results show that the mean square error performance of the improved algorithm in the channel with unknown sparsity is 4dB higher than that of the original algorithm.","PeriodicalId":342968,"journal":{"name":"2021 2nd International Symposium on Computer Engineering and Intelligent Communications (ISCEIC)","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116963553","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 : 2021-08-01DOI: 10.1109/ISCEIC53685.2021.00026
Zhaolong Gao, Mengvi P. Gatpandan, Paulino H. Gatpandan
With the continuous development of machine learning, more and more applications are applied to all areas of life. Decision tree algorithm, as a classic algorithm in machine learning, is also widely used in various industries. In recent years, the informatization of the college entrance examination has allowed the admissions department to accumulate a large amount of college entrance examination data. Based on the decision tree algorithm in machine learning, the candidate data is analyzed, and an algorithm for predicting candidates to apply for the major is proposed, and the admission data of a certain university in Yantai is used as an experiment. The data set tests the prediction accuracy of the algorithm. Two complete decision trees have been constructed to provide high school seniors with intelligent decisions and suggestions for a series of basic issues such as college selection, major selection, and voluntary reporting.
{"title":"Classification Decision Tree Algorithm in Predicting Students’ Course Preference","authors":"Zhaolong Gao, Mengvi P. Gatpandan, Paulino H. Gatpandan","doi":"10.1109/ISCEIC53685.2021.00026","DOIUrl":"https://doi.org/10.1109/ISCEIC53685.2021.00026","url":null,"abstract":"With the continuous development of machine learning, more and more applications are applied to all areas of life. Decision tree algorithm, as a classic algorithm in machine learning, is also widely used in various industries. In recent years, the informatization of the college entrance examination has allowed the admissions department to accumulate a large amount of college entrance examination data. Based on the decision tree algorithm in machine learning, the candidate data is analyzed, and an algorithm for predicting candidates to apply for the major is proposed, and the admission data of a certain university in Yantai is used as an experiment. The data set tests the prediction accuracy of the algorithm. Two complete decision trees have been constructed to provide high school seniors with intelligent decisions and suggestions for a series of basic issues such as college selection, major selection, and voluntary reporting.","PeriodicalId":342968,"journal":{"name":"2021 2nd International Symposium on Computer Engineering and Intelligent Communications (ISCEIC)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123631318","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 : 2021-08-01DOI: 10.1109/ISCEIC53685.2021.00033
Xiaodong Xie, Bin Qin, Ziyun Wan, Wei Nie
Text sentiment analysis is an important research topic in the field of natural language processing. Compared with coarse-grained text sentiment analysis with a single sentiment tendency based on sentences or documents, fine-grained aspect- level sentiment analysis is more suitable for practical application scenarios and becomes more difficult at the same time. In this paper, we propose a multi-task-based joint learning model for aspect-level sentiment analysis of text, using the BERT_CBiGRU composite network with stronger text semantic representation as the main learning task for aspect-level sentiment analysis, and designing an auxiliary learning task for aspect target identification. For the sample imbalance problem, we introduce the focal loss function Focal Loss. Final experiments show that our model has a certain improvement compared with previous models.
{"title":"Text Aspect-level Sentiment Analysis based on Multi- task Joint Learning","authors":"Xiaodong Xie, Bin Qin, Ziyun Wan, Wei Nie","doi":"10.1109/ISCEIC53685.2021.00033","DOIUrl":"https://doi.org/10.1109/ISCEIC53685.2021.00033","url":null,"abstract":"Text sentiment analysis is an important research topic in the field of natural language processing. Compared with coarse-grained text sentiment analysis with a single sentiment tendency based on sentences or documents, fine-grained aspect- level sentiment analysis is more suitable for practical application scenarios and becomes more difficult at the same time. In this paper, we propose a multi-task-based joint learning model for aspect-level sentiment analysis of text, using the BERT_CBiGRU composite network with stronger text semantic representation as the main learning task for aspect-level sentiment analysis, and designing an auxiliary learning task for aspect target identification. For the sample imbalance problem, we introduce the focal loss function Focal Loss. Final experiments show that our model has a certain improvement compared with previous models.","PeriodicalId":342968,"journal":{"name":"2021 2nd International Symposium on Computer Engineering and Intelligent Communications (ISCEIC)","volume":"1120 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129459272","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 : 2021-08-01DOI: 10.1109/ISCEIC53685.2021.00017
Qi Wan, Bingqing Han
Mobility is the most distinctive and important feature of Mobile Wireless Ad Hoc Networks(MANET), which brings many limitations along with its wide applications. A good mobility model is a good basis for further improving the performance of routing layer and MAC layer. Therefore, a movement strategy is proposed for the node movement problem. Only those nodes that are poorly distributed and meet the conditions are selected to form the set of mobile nodes. Neighborhood distribution information and residual energy information are fully considered for improving the initial population and fitness function of the genetic algorithm. Then, the improved genetic algorithm is used to search for the optimal next mobile position for each node in the set. The simulation results show that the nodes in the network can achieve higher network coverage and connectivity with fewer moves, regardless of whether the nodes are well or poorly distributed.
{"title":"Mobile Strategy for Nodes in Mobile Wireless Ad Hoc Networks Based on Genetic Algorithm","authors":"Qi Wan, Bingqing Han","doi":"10.1109/ISCEIC53685.2021.00017","DOIUrl":"https://doi.org/10.1109/ISCEIC53685.2021.00017","url":null,"abstract":"Mobility is the most distinctive and important feature of Mobile Wireless Ad Hoc Networks(MANET), which brings many limitations along with its wide applications. A good mobility model is a good basis for further improving the performance of routing layer and MAC layer. Therefore, a movement strategy is proposed for the node movement problem. Only those nodes that are poorly distributed and meet the conditions are selected to form the set of mobile nodes. Neighborhood distribution information and residual energy information are fully considered for improving the initial population and fitness function of the genetic algorithm. Then, the improved genetic algorithm is used to search for the optimal next mobile position for each node in the set. The simulation results show that the nodes in the network can achieve higher network coverage and connectivity with fewer moves, regardless of whether the nodes are well or poorly distributed.","PeriodicalId":342968,"journal":{"name":"2021 2nd International Symposium on Computer Engineering and Intelligent Communications (ISCEIC)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129938529","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 : 2021-08-01DOI: 10.1109/ISCEIC53685.2021.00016
Yunfeng Zhu
Attitude calculation is a key technology used in many fields. Reliability and stability of attitude calculation are highly required in industrial design, which needs to adopt multi-sensor data fusion technology. The traditional complementary filtering algorithm has the problems of slow attenuation speed of stopband and steady-state error caused by constant drift of gyroscope. An improved complementary filtering algorithm based on quaternion is proposed for attitude calculation. By adding an integral link on the basis of complementary filtering, complementary filtering algorithm can make better use of accelerometers and magnetometers to compensate gyroscopes, reduce the steady-state error of the system and improve the stability of attitude solution System. An experimental platform based on STM32F103 was built for verification. The experimental results show, under static conditions, the improved complementary filtering algorithm is compared with the traditional complementary filtering algorithm, the stability of roll angle, pitch angle and heading angle are increased by 36.24%,32.0% and 18.0% respectively. Under dynamic conditions, the posture angle fluctuations calculated by the improved algorithm are smaller, and stabilization and convergence are faster. Compared to a conventional complementary filtering algorithm is controlled more easily.
{"title":"Posture Calculation of Improved Complementary Filtering Algorithm Based on Quaternion","authors":"Yunfeng Zhu","doi":"10.1109/ISCEIC53685.2021.00016","DOIUrl":"https://doi.org/10.1109/ISCEIC53685.2021.00016","url":null,"abstract":"Attitude calculation is a key technology used in many fields. Reliability and stability of attitude calculation are highly required in industrial design, which needs to adopt multi-sensor data fusion technology. The traditional complementary filtering algorithm has the problems of slow attenuation speed of stopband and steady-state error caused by constant drift of gyroscope. An improved complementary filtering algorithm based on quaternion is proposed for attitude calculation. By adding an integral link on the basis of complementary filtering, complementary filtering algorithm can make better use of accelerometers and magnetometers to compensate gyroscopes, reduce the steady-state error of the system and improve the stability of attitude solution System. An experimental platform based on STM32F103 was built for verification. The experimental results show, under static conditions, the improved complementary filtering algorithm is compared with the traditional complementary filtering algorithm, the stability of roll angle, pitch angle and heading angle are increased by 36.24%,32.0% and 18.0% respectively. Under dynamic conditions, the posture angle fluctuations calculated by the improved algorithm are smaller, and stabilization and convergence are faster. Compared to a conventional complementary filtering algorithm is controlled more easily.","PeriodicalId":342968,"journal":{"name":"2021 2nd International Symposium on Computer Engineering and Intelligent Communications (ISCEIC)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127802847","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 : 2021-08-01DOI: 10.1109/ISCEIC53685.2021.00034
Yue Sun, Sen Jing, Ranxin Gao, Bin Dong, Wei Chen, Xiaotian Xu
In order to improve the security protection ability and service quality of web application system, this paper focused on the analysis of attack types and characteristics of typical web application and studied the identification and detection technology of attacks logs with eigenvalue and semantic analysis. Finally, a multi-source aggregation analysis technology architecture of attacks logs for web application is proposed. Based on association analysis strategy, this architecture can effectively identify the attack behavior, predict the attack situation, and effectively improve the ability in resisting external threats of web application.
{"title":"Research on the Detection and Analysis Technology in Web Application Attacks Logs","authors":"Yue Sun, Sen Jing, Ranxin Gao, Bin Dong, Wei Chen, Xiaotian Xu","doi":"10.1109/ISCEIC53685.2021.00034","DOIUrl":"https://doi.org/10.1109/ISCEIC53685.2021.00034","url":null,"abstract":"In order to improve the security protection ability and service quality of web application system, this paper focused on the analysis of attack types and characteristics of typical web application and studied the identification and detection technology of attacks logs with eigenvalue and semantic analysis. Finally, a multi-source aggregation analysis technology architecture of attacks logs for web application is proposed. Based on association analysis strategy, this architecture can effectively identify the attack behavior, predict the attack situation, and effectively improve the ability in resisting external threats of web application.","PeriodicalId":342968,"journal":{"name":"2021 2nd International Symposium on Computer Engineering and Intelligent Communications (ISCEIC)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126742155","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 : 2021-08-01DOI: 10.1109/ISCEIC53685.2021.00077
Yue Sun, Bin Dong, Wei Chen, Xiaotian Xu, Guanlin Si, Sen Jing
The application of intelligent video terminal has spread in all aspects of production and life, such as urban transportation, enterprises, hospitals, banks, and families. In recent years, intelligent video terminals, video recorders and other video monitoring system components are frequently exposed to high risks of security vulnerabilities, which is likely to threaten the privacy of users and data security. Therefore, it is necessary to strengthen the security research and testing of intelligent video terminals, and formulate reinforcement and protection strategies based on the evaluation results, in order to ensure the confidentiality, integrity and availability of data collected and transmitted by intelligent video terminals.
{"title":"Research on Security Evaluation Technology of Intelligent Video Terminal","authors":"Yue Sun, Bin Dong, Wei Chen, Xiaotian Xu, Guanlin Si, Sen Jing","doi":"10.1109/ISCEIC53685.2021.00077","DOIUrl":"https://doi.org/10.1109/ISCEIC53685.2021.00077","url":null,"abstract":"The application of intelligent video terminal has spread in all aspects of production and life, such as urban transportation, enterprises, hospitals, banks, and families. In recent years, intelligent video terminals, video recorders and other video monitoring system components are frequently exposed to high risks of security vulnerabilities, which is likely to threaten the privacy of users and data security. Therefore, it is necessary to strengthen the security research and testing of intelligent video terminals, and formulate reinforcement and protection strategies based on the evaluation results, in order to ensure the confidentiality, integrity and availability of data collected and transmitted by intelligent video terminals.","PeriodicalId":342968,"journal":{"name":"2021 2nd International Symposium on Computer Engineering and Intelligent Communications (ISCEIC)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126750143","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 : 2021-08-01DOI: 10.1109/ISCEIC53685.2021.00056
Zekun Huang, Minzheng Li
With the rapid development of smart grids, inspection and maintenance of power equipment is essential. The large bandwidth, high reliability and low latency communication of 5G technology can meet the needs of drone inspections. However, the environment of the transmission line corridor is complex and changeable. During the UAV inspection, the channel environment and the channel transmission coefficient between the UAV and the base station change in real time. In order to ensure the reliability and effectiveness of communication, the UAV terminal needs to estimate and track the changed channel transmission coefficient in real time, and feed it back to the base station. Based on the Massive-MIMO millimeter wave drone communication scenario, this paper constructs the channel model of the drone terminal and the base station in this scenario and solves the estimation, trackingand and feedback of the channel state information in this scenario by fusing deep learning algorithms problem.
{"title":"Research on Channel Feedback Algorithm in UAV Inspection Communication Subsystem of Smart Grid","authors":"Zekun Huang, Minzheng Li","doi":"10.1109/ISCEIC53685.2021.00056","DOIUrl":"https://doi.org/10.1109/ISCEIC53685.2021.00056","url":null,"abstract":"With the rapid development of smart grids, inspection and maintenance of power equipment is essential. The large bandwidth, high reliability and low latency communication of 5G technology can meet the needs of drone inspections. However, the environment of the transmission line corridor is complex and changeable. During the UAV inspection, the channel environment and the channel transmission coefficient between the UAV and the base station change in real time. In order to ensure the reliability and effectiveness of communication, the UAV terminal needs to estimate and track the changed channel transmission coefficient in real time, and feed it back to the base station. Based on the Massive-MIMO millimeter wave drone communication scenario, this paper constructs the channel model of the drone terminal and the base station in this scenario and solves the estimation, trackingand and feedback of the channel state information in this scenario by fusing deep learning algorithms problem.","PeriodicalId":342968,"journal":{"name":"2021 2nd International Symposium on Computer Engineering and Intelligent Communications (ISCEIC)","volume":"115 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132847015","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 : 2021-08-01DOI: 10.1109/ISCEIC53685.2021.00019
Si Chen, Fang Lv, Ping Huo
In order to improve the timeliness and accuracy of the automatic detection and recognition of sunflower leaf diseases in Inner Mongolia, this study used an improved Yolov4 model to detect and recognize the three diseases of sunflower leaf verticillium wilt, powdery mildew and rust. First, use the K-means clustering algorithm to cluster sunflower disease samples to generate a new anchor box size, making the anchor box of the Yolov4 network model more suitable for the detection of smaller target lesions on sunflower leaves; secondly, define the three network functions of MobileNetV1, MobileNetV2 and MobileNetV3, obtain three effective feature layers corresponding to each MobileNet network, and use these three effective feature layers to replace the effective feature layer of the original Yolov4 backbone network CSP-Darknet53 as the backbone feature extraction network. The experimental results show that the accuracy and recall of the improved method proposed in this paper are significantly improved compared with the original Yolov4 model. Therefore, the deep learning model trained by the improved Yolov4 algorithm has better robustness and is more suitable for sunflower leaf disease detection.
{"title":"Improved detection of yolov4 sunflower leaf diseases","authors":"Si Chen, Fang Lv, Ping Huo","doi":"10.1109/ISCEIC53685.2021.00019","DOIUrl":"https://doi.org/10.1109/ISCEIC53685.2021.00019","url":null,"abstract":"In order to improve the timeliness and accuracy of the automatic detection and recognition of sunflower leaf diseases in Inner Mongolia, this study used an improved Yolov4 model to detect and recognize the three diseases of sunflower leaf verticillium wilt, powdery mildew and rust. First, use the K-means clustering algorithm to cluster sunflower disease samples to generate a new anchor box size, making the anchor box of the Yolov4 network model more suitable for the detection of smaller target lesions on sunflower leaves; secondly, define the three network functions of MobileNetV1, MobileNetV2 and MobileNetV3, obtain three effective feature layers corresponding to each MobileNet network, and use these three effective feature layers to replace the effective feature layer of the original Yolov4 backbone network CSP-Darknet53 as the backbone feature extraction network. The experimental results show that the accuracy and recall of the improved method proposed in this paper are significantly improved compared with the original Yolov4 model. Therefore, the deep learning model trained by the improved Yolov4 algorithm has better robustness and is more suitable for sunflower leaf disease detection.","PeriodicalId":342968,"journal":{"name":"2021 2nd International Symposium on Computer Engineering and Intelligent Communications (ISCEIC)","volume":"23 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130995153","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：