Based on the perspective of Sustainable Development Goals (SDGs) Quality Education and lifelong learning, it is necessary to respect the learning opportunities and quality for all individuals. Online learning can provide more opportunities for lifelong learning, but due to the significant differences in students’ backgrounds and characteristics, personalized and timely support becomes more crucial. Learning analytics (LA) in online learning environment is a way to facilitate understanding of the potential meaningful information and relationships of students. One of the main functions of LA is to monitor the learning performance and identify potential learning problems early. In this study, 𝑘-means clustering is performed to determine the types of learning in lifelong online learning environments, based on students’ personal traits (background factors), learning behavior paths, and interactive perspectives on learning performance. Moreover, statistical analysis is used to further evaluate the linear correlation coefficients as well as the characteristics of each group of students, who ranged in age from 18 to 73, with a total of 2386 participants from five courses, in the interactive perspective. The result shows a significant correlation between learning performance and persistence across the three learning clusters, with a tendency towards continuous learning, thus providing educators an understanding of the learning behavior characteristics of those types of online learners.
{"title":"An Analysis of Online Learner Types Applicable to Lifelong Learning Environments","authors":"Shiow-Lin Hwu Shiow-Lin Hwu, Sheng-Lung Peng Shiow-Lin Hwu","doi":"10.53106/199115992023083404020","DOIUrl":"https://doi.org/10.53106/199115992023083404020","url":null,"abstract":"\u0000 Based on the perspective of Sustainable Development Goals (SDGs) Quality Education and lifelong learning, it is necessary to respect the learning opportunities and quality for all individuals. Online learning can provide more opportunities for lifelong learning, but due to the significant differences in students’ backgrounds and characteristics, personalized and timely support becomes more crucial. Learning analytics (LA) in online learning environment is a way to facilitate understanding of the potential meaningful information and relationships of students. One of the main functions of LA is to monitor the learning performance and identify potential learning problems early. In this study, 𝑘-means clustering is performed to determine the types of learning in lifelong online learning environments, based on students’ personal traits (background factors), learning behavior paths, and interactive perspectives on learning performance. Moreover, statistical analysis is used to further evaluate the linear correlation coefficients as well as the characteristics of each group of students, who ranged in age from 18 to 73, with a total of 2386 participants from five courses, in the interactive perspective. The result shows a significant correlation between learning performance and persistence across the three learning clusters, with a tendency towards continuous learning, thus providing educators an understanding of the learning behavior characteristics of those types of online learners.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124343917","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 : 2023-08-01DOI: 10.53106/199115992023083404003
Jin-Ze Du Jin-Ze Du, Jun-Wei Liu Jin-Ze Du, Tao Feng Jun-Wei Liu, Zhan-Ting Yuan Tao Feng
In order to verify the security of the Z-Wave communication protocol, the possible attacks in the protocol are analyzed to reduce user privacy security vulnerabilities. For the communication process and key exchange process between the controller and the node, this paper uses CPN tools to model the Z-Wave S2 protocol, and introduces the Dolev-Yao attack model to verify the security behavior of the protocol. The results show that there is a man-in-the-middle attack when using S2 authentication for device inclusion. In response to this vulnerability, we propose a lightweight static authentication scheme based on HKDF function and XOR operation, which performs authentication between Z-Wave controller and slave device. Secondly, we formally verify the security objectives of the improved scheme, and prove that the optimization scheme can effectively prevent man-in-the-middle attacks in the S2 security mode.
{"title":"Formal Analysis and Improvement of Z-Wave Protocol","authors":"Jin-Ze Du Jin-Ze Du, Jun-Wei Liu Jin-Ze Du, Tao Feng Jun-Wei Liu, Zhan-Ting Yuan Tao Feng","doi":"10.53106/199115992023083404003","DOIUrl":"https://doi.org/10.53106/199115992023083404003","url":null,"abstract":"\u0000 In order to verify the security of the Z-Wave communication protocol, the possible attacks in the protocol are analyzed to reduce user privacy security vulnerabilities. For the communication process and key exchange process between the controller and the node, this paper uses CPN tools to model the Z-Wave S2 protocol, and introduces the Dolev-Yao attack model to verify the security behavior of the protocol. The results show that there is a man-in-the-middle attack when using S2 authentication for device inclusion. In response to this vulnerability, we propose a lightweight static authentication scheme based on HKDF function and XOR operation, which performs authentication between Z-Wave controller and slave device. Secondly, we formally verify the security objectives of the improved scheme, and prove that the optimization scheme can effectively prevent man-in-the-middle attacks in the S2 security mode. \u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132676430","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 : 2023-08-01DOI: 10.53106/199115992023083404012
Peng-Jie Du Peng-Jie Du, Mu-Zhuo Zhang Peng-Jie Du
In order to solve the technical requirements for automatic recognition and judgment of pantograph wear degree of Multiple Units, this paper designs a network structure based on Mask R-CNN structure. At the same time, in order to improve the ability of image feature extraction in the network, the original backbone network is replaced with ResNet-50, a residual network with more prominent feature extraction ability. Secondly, in order to improve the ability to search for targets in the image, the detection head is reconstructed, to improve the recognition ability of targets. Finally, the effectiveness of the algorithm and its ability to identify pantograph faults were verified through simulation experiments.
{"title":"Computer Vision Aided Pantograph Fault Identification Method for Multiple Units","authors":"Peng-Jie Du Peng-Jie Du, Mu-Zhuo Zhang Peng-Jie Du","doi":"10.53106/199115992023083404012","DOIUrl":"https://doi.org/10.53106/199115992023083404012","url":null,"abstract":"\u0000 In order to solve the technical requirements for automatic recognition and judgment of pantograph wear degree of Multiple Units, this paper designs a network structure based on Mask R-CNN structure. At the same time, in order to improve the ability of image feature extraction in the network, the original backbone network is replaced with ResNet-50, a residual network with more prominent feature extraction ability. Secondly, in order to improve the ability to search for targets in the image, the detection head is reconstructed, to improve the recognition ability of targets. Finally, the effectiveness of the algorithm and its ability to identify pantograph faults were verified through simulation experiments.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123997306","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 : 2023-08-01DOI: 10.53106/199115992023083404009
dan Zhang dan ZHANG, Ting-Jie Lu Dan Zhang, Chen-Xing Yang Ting-Jie Lu, Xue-Yan Wang Chen-Xing Yang
In this paper, the principal component analysis method is mainly used to analyze that the construction level of Smart Court in different regions. The construction level of Smart Court consists of three main components. The result of Smart Court construction level in different regions is calculated. It is concluded that the construction level of Smart Court varies in different regions. It is necessary to make great efforts to solve the problem of unbalanced regional development. It is suggested that the construction of Smart Court in the future should focus on the publicity business such as online cases. And more and more new generation information technology such as artificial intelligence should be applied.
{"title":"Structure and Analysis of the Smart Court Construction Evaluation Index Based on Principal Component Analysis for Different Regions","authors":"dan Zhang dan ZHANG, Ting-Jie Lu Dan Zhang, Chen-Xing Yang Ting-Jie Lu, Xue-Yan Wang Chen-Xing Yang","doi":"10.53106/199115992023083404009","DOIUrl":"https://doi.org/10.53106/199115992023083404009","url":null,"abstract":"\u0000 In this paper, the principal component analysis method is mainly used to analyze that the construction level of Smart Court in different regions. The construction level of Smart Court consists of three main components. The result of Smart Court construction level in different regions is calculated. It is concluded that the construction level of Smart Court varies in different regions. It is necessary to make great efforts to solve the problem of unbalanced regional development. It is suggested that the construction of Smart Court in the future should focus on the publicity business such as online cases. And more and more new generation information technology such as artificial intelligence should be applied.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125820724","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 : 2023-08-01DOI: 10.53106/199115992023083404019
Chaoqun Kang Chaoqun Kang, Erxia Li Chaoqun Kang, Dongxiao Liu Erxia Li, Xinhong You Dongxiao Liu, Xiaoyong Li Xinhong You
With the diversity and complexity of user access behaviors in the “micro-segmentation” cloud computing environment, it is no longer possible to control unauthorized access of authorized users by only relying on user identity login authentication to control user access to cloud resources. The existing trust evaluation methods can not cope with the characteristics of “micro-isolated” cloud environment, which is characterized by high granularity of resources, increasing number of users’ access requests and rapid changes. Based on the zero-trust principle of “Never trust, al-ways verify”, we propose a dynamic, fine-grained user trust evaluation model for micro-segmentation cloud computing environment, which combines multiple user trust attributes and leverages the subjective-objective approach to assign weights to trust attribute indicators to achieve dynamic scoring of users’ real-time behaviors. To capture the characteristics of users’ intrinsic behaviors, we use correlation analysis to identify the correlation between users’ current and historical behaviors, and combine sliding windows and penalty functions to optimize the model. The massive simulation experiments demonstrate the effectiveness of the proposed dynamic and fine-grained method, which can effectively combine the intrinsic correlation of users’ own access behavior and the difference of access behavior among different users.
{"title":"A Dynamic and Fine-Grained User Trust Evaluation Model for Micro-Segmentation Cloud Computing Environment","authors":"Chaoqun Kang Chaoqun Kang, Erxia Li Chaoqun Kang, Dongxiao Liu Erxia Li, Xinhong You Dongxiao Liu, Xiaoyong Li Xinhong You","doi":"10.53106/199115992023083404019","DOIUrl":"https://doi.org/10.53106/199115992023083404019","url":null,"abstract":"\u0000 With the diversity and complexity of user access behaviors in the “micro-segmentation” cloud computing environment, it is no longer possible to control unauthorized access of authorized users by only relying on user identity login authentication to control user access to cloud resources. The existing trust evaluation methods can not cope with the characteristics of “micro-isolated” cloud environment, which is characterized by high granularity of resources, increasing number of users’ access requests and rapid changes. Based on the zero-trust principle of “Never trust, al-ways verify”, we propose a dynamic, fine-grained user trust evaluation model for micro-segmentation cloud computing environment, which combines multiple user trust attributes and leverages the subjective-objective approach to assign weights to trust attribute indicators to achieve dynamic scoring of users’ real-time behaviors. To capture the characteristics of users’ intrinsic behaviors, we use correlation analysis to identify the correlation between users’ current and historical behaviors, and combine sliding windows and penalty functions to optimize the model. The massive simulation experiments demonstrate the effectiveness of the proposed dynamic and fine-grained method, which can effectively combine the intrinsic correlation of users’ own access behavior and the difference of access behavior among different users.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121422717","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 : 2023-08-01DOI: 10.53106/199115992023083404002
Hong-Rong Jing Hong-Rong Jing, Guo-Jun Lin Hong-Rong Jing, Zhong-Ling Liu Guo-Jun Lin, Jing-Li Zhong-Ling Liu, Jing-Li He Li, Xuan-Han Li Li He, Hong-Jie Zhang Xuan-Han Li, Shun-Yong Zhou Hong-Jie Zhang
We develop a more efficient lightweight network based on SE-ShuffleNet V2 to address the issues of large parameter sizes and sluggish feature extraction rates in large networks in the field of face recognition. First, to increase the network’s accuracy and inference speed, the ReLU activation function of the original ShuffleNet V2 basic unit is swapped out for a segmented linear activation function. Second, the SE attention mechanism is added to the lightweight network ShuffleNet V2, which may improve the effective feature weights and decrease the invalid feature weights, and the SE attention causes the network to focus on more helpful features. Finally, the addition of the Arcface loss function enhances the face recognition network’s capacity for categorization. Experiments indicate that the SE-ShuffleNet V2 network that we created achieves superior performance under the parameters of position and age. Particularly, the LFW accuracy is 99.38%. The algorithm presented in this research significantly increases face recognition accuracy when compared to the original ShuffleNet V2 network, therefore the additional parameters and longer inference times can be disregarded. To match the accuracy of substantial convolutional networks, we developed the lightweight SE-ShuffleNet V2.
{"title":"Face Recognition Method Based on Lightweight Network SE-ShuffleNet V2","authors":"Hong-Rong Jing Hong-Rong Jing, Guo-Jun Lin Hong-Rong Jing, Zhong-Ling Liu Guo-Jun Lin, Jing-Li Zhong-Ling Liu, Jing-Li He Li, Xuan-Han Li Li He, Hong-Jie Zhang Xuan-Han Li, Shun-Yong Zhou Hong-Jie Zhang","doi":"10.53106/199115992023083404002","DOIUrl":"https://doi.org/10.53106/199115992023083404002","url":null,"abstract":"\u0000 We develop a more efficient lightweight network based on SE-ShuffleNet V2 to address the issues of large parameter sizes and sluggish feature extraction rates in large networks in the field of face recognition. First, to increase the network’s accuracy and inference speed, the ReLU activation function of the original ShuffleNet V2 basic unit is swapped out for a segmented linear activation function. Second, the SE attention mechanism is added to the lightweight network ShuffleNet V2, which may improve the effective feature weights and decrease the invalid feature weights, and the SE attention causes the network to focus on more helpful features. Finally, the addition of the Arcface loss function enhances the face recognition network’s capacity for categorization. Experiments indicate that the SE-ShuffleNet V2 network that we created achieves superior performance under the parameters of position and age. Particularly, the LFW accuracy is 99.38%. The algorithm presented in this research significantly increases face recognition accuracy when compared to the original ShuffleNet V2 network, therefore the additional parameters and longer inference times can be disregarded. To match the accuracy of substantial convolutional networks, we developed the lightweight SE-ShuffleNet V2.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115023139","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}
In this study, we investigate the application of a deep learning framework for the recognition of pig vocalizations. This innovative approach aims to actively monitor and evaluate the diverse states of pigs, with an overarching objective to improve the efficiency of pig farming through prompt identification and resolution of issues. In our comprehensive data collection effort, we carefully gathered a vast assortment of vocal samples from 50 pigs, representative of four distinct states: normal, frightened, coughing, and sneezing. We then meticulously analyzed this vocal data using Mel Frequency Cepstral Coefficients (MFCC). For accurate recognition of pig vocalizations, we devised a fusion model that combines the strengths of Residual Networks (ResNet) and Long Short-Term Memory Networks (LSTM). This model was subsequently tailored, trained, and optimized to meet our specific requirements. Upon rigorous evaluation, we found our model to exhibit exceptional performance in pig vocal recognition tasks, thereby reinforcing the potential of deep learning methodologies in revolutionizing the livestock industry. This research notably underscores the potential of deploying efficient real-time health monitoring systems, offering a promising avenue towards modernizing livestock management practices.
{"title":"Animal Vocal Recognition-Based Breeding Tracking and Disease Warning","authors":"Yinggang Xie Yinggang Xie, Yangpeng Xiao Yinggang Xie, Xuewei Peng Yangpeng Xiao, Qijia Liu Xuewei Peng","doi":"10.53106/199115992023083404011","DOIUrl":"https://doi.org/10.53106/199115992023083404011","url":null,"abstract":"\u0000 In this study, we investigate the application of a deep learning framework for the recognition of pig vocalizations. This innovative approach aims to actively monitor and evaluate the diverse states of pigs, with an overarching objective to improve the efficiency of pig farming through prompt identification and resolution of issues. In our comprehensive data collection effort, we carefully gathered a vast assortment of vocal samples from 50 pigs, representative of four distinct states: normal, frightened, coughing, and sneezing. We then meticulously analyzed this vocal data using Mel Frequency Cepstral Coefficients (MFCC). For accurate recognition of pig vocalizations, we devised a fusion model that combines the strengths of Residual Networks (ResNet) and Long Short-Term Memory Networks (LSTM). This model was subsequently tailored, trained, and optimized to meet our specific requirements. Upon rigorous evaluation, we found our model to exhibit exceptional performance in pig vocal recognition tasks, thereby reinforcing the potential of deep learning methodologies in revolutionizing the livestock industry. This research notably underscores the potential of deploying efficient real-time health monitoring systems, offering a promising avenue towards modernizing livestock management practices.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127660425","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 : 2023-08-01DOI: 10.53106/199115992023083404015
Xiaoning Bo Xiaoning Bo, Jin Wang Xiaoning Bo, Honglan Li Jin Wang, Guoqin Li Honglan Li, Feng Lu Guoqin Li
This paper proposes a machine vision based solder pad detection method to improve the detection accuracy and efficiency of PCB solder pad defects in electronic components due to missed detection and low detection efficiency. Firstly, preprocess the electronic component pad images collected by the visual system, then use threshold segmentation method to perform preliminary segmentation of the pad images. Then, the coarse segmented images are finely segmented using mean clustering method, and the fine segmented images are pixel edge extracted. Finally, the matrix subpixel edge detection method is used to improve the edge detection accuracy. Simulation experiments have shown that the proposed method can significantly improve the accuracy and speed of defect recognition.
{"title":"Machine Vision Based Defect Detection Method for Electronic Component Solder Pads","authors":"Xiaoning Bo Xiaoning Bo, Jin Wang Xiaoning Bo, Honglan Li Jin Wang, Guoqin Li Honglan Li, Feng Lu Guoqin Li","doi":"10.53106/199115992023083404015","DOIUrl":"https://doi.org/10.53106/199115992023083404015","url":null,"abstract":"\u0000 This paper proposes a machine vision based solder pad detection method to improve the detection accuracy and efficiency of PCB solder pad defects in electronic components due to missed detection and low detection efficiency. Firstly, preprocess the electronic component pad images collected by the visual system, then use threshold segmentation method to perform preliminary segmentation of the pad images. Then, the coarse segmented images are finely segmented using mean clustering method, and the fine segmented images are pixel edge extracted. Finally, the matrix subpixel edge detection method is used to improve the edge detection accuracy. Simulation experiments have shown that the proposed method can significantly improve the accuracy and speed of defect recognition.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116655811","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 : 2023-06-01DOI: 10.53106/199115992023063403012
Yonggang Dong Yonggang Dong, Shichao Cao Yonggang Dong, Haoliang Yang Shichao Cao
A large number of distributed energy sources connected to the grid will cause certain disturbances to the stability of the grid system. We consider the random characteristics and establish a dynamic simulation framework for the active power distribution system suitable for the integral projection algorithm. The article uses an internal integrator to solve the fast dynamic process with explicit and implicit Euler methods in small steps. In the calculation process, this method can effectively consider the influence of events such as fault disturbance on the grid connection of the distributed power grid. Numerical analysis and simulation tests verify the effectiveness of this algorithm.
{"title":"Research on Influence of Distributed Energy Accessing Power Grid System Based on Integral Projection Algorithm","authors":"Yonggang Dong Yonggang Dong, Shichao Cao Yonggang Dong, Haoliang Yang Shichao Cao","doi":"10.53106/199115992023063403012","DOIUrl":"https://doi.org/10.53106/199115992023063403012","url":null,"abstract":"\u0000 A large number of distributed energy sources connected to the grid will cause certain disturbances to the stability of the grid system. We consider the random characteristics and establish a dynamic simulation framework for the active power distribution system suitable for the integral projection algorithm. The article uses an internal integrator to solve the fast dynamic process with explicit and implicit Euler methods in small steps. In the calculation process, this method can effectively consider the influence of events such as fault disturbance on the grid connection of the distributed power grid. Numerical analysis and simulation tests verify the effectiveness of this algorithm.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127456215","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}
As a popular research direction in the field of intelligent transportation, various scholars have widely concerned themselves with traffic sign detection However, there are still some key issues that need to be further solved in order to thoroughly apply related technologies to real scenarios, such as the feature extraction scheme of traffic sign images, and the optimal selection of detection methods. For the purpose of overcoming these difficulties. This paper proposes a YOLO-based traffic sign detection framework. Firstly, a lightweight convolution attention mechanism is embedded into the backbone network to obtain the information of space and channel; Secondly, the multi-scale awareness module is used to replace large convolution with 3×3 convolution superposition to improve the receptive field area of the object in the model and enhance the feature fusion performance of the model; Finally, CIoU is used as the loss function of the bounding box to locate the experimental object with high precision. The experimental results show that on the CCTSDB data set, the MAP of this method reaches 91.0%, which is 3.5% higher than the original YOLOv5. Compared with other mainstream object detection algorithms, it has a certain degree of improvement, which proves the effectiveness of this method.
{"title":"Traffic Sign Detection Based on Improved YOLOv5","authors":"Hua-Ping Zhou Hua-Ping Zhou, Chen-Chen Xu Hua-Ping Zhou, Ke-Lei Sun Chen-Chen Xu","doi":"10.53106/199115992023063403005","DOIUrl":"https://doi.org/10.53106/199115992023063403005","url":null,"abstract":"\u0000 As a popular research direction in the field of intelligent transportation, various scholars have widely concerned themselves with traffic sign detection However, there are still some key issues that need to be further solved in order to thoroughly apply related technologies to real scenarios, such as the feature extraction scheme of traffic sign images, and the optimal selection of detection methods. For the purpose of overcoming these difficulties. This paper proposes a YOLO-based traffic sign detection framework. Firstly, a lightweight convolution attention mechanism is embedded into the backbone network to obtain the information of space and channel; Secondly, the multi-scale awareness module is used to replace large convolution with 3×3 convolution superposition to improve the receptive field area of the object in the model and enhance the feature fusion performance of the model; Finally, CIoU is used as the loss function of the bounding box to locate the experimental object with high precision. The experimental results show that on the CCTSDB data set, the MAP of this method reaches 91.0%, which is 3.5% higher than the original YOLOv5. Compared with other mainstream object detection algorithms, it has a certain degree of improvement, which proves the effectiveness of this method.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125509414","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：