Pub Date : 2024-01-01DOI: 10.53106/160792642024012501003
Hong-Yi Pai Hong-Yi Pai, Ching-Huang Wang Hong-Yi Pai, Yi-Chen Lai Ching-Huang Wang
In this paper, we propose Virtual Avatar Interactive Live Streaming System (VAILSS), our approach to streaming 3D avatar performances in interactive storytelling applications using multimodal interactive techniques. This study focuses on developing a virtual live interaction system for VTubers in the context of language learning. A live virtual performance from our system consists of three components: 1) an Avatar generation framework, 2) an AI motion capture system, and 3) an interactive storytelling engine. The system integrates artificial intelligence and uses motion capture to identify facial expressions and movements, enabling virtual characters to provide live storytelling services. Additionally, the system allows for bi-directional interaction with players through tablet touch and voice. The system aims to promote multimodal learning channels for the masses. To gain deeper insights into the effectiveness of virtual role-playing in language learning, we organized a three-week workshop to investigate our system’s impact on user experience. We extended invitations to 17 night-school freshmen from the NFU Department of Applied Foreign Languages, enrolled in an English class focused on tour guides, to participate in a virtual performance. The experiment results demonstrated that the VAILSS positively affected students’ learning outcomes, particularly in enhancing foreign language acquisition.
{"title":"Development of an Interactive Live Streaming System for Language Learning","authors":"Hong-Yi Pai Hong-Yi Pai, Ching-Huang Wang Hong-Yi Pai, Yi-Chen Lai Ching-Huang Wang","doi":"10.53106/160792642024012501003","DOIUrl":"https://doi.org/10.53106/160792642024012501003","url":null,"abstract":"\u0000 In this paper, we propose Virtual Avatar Interactive Live Streaming System (VAILSS), our approach to streaming 3D avatar performances in interactive storytelling applications using multimodal interactive techniques. This study focuses on developing a virtual live interaction system for VTubers in the context of language learning. A live virtual performance from our system consists of three components: 1) an Avatar generation framework, 2) an AI motion capture system, and 3) an interactive storytelling engine. The system integrates artificial intelligence and uses motion capture to identify facial expressions and movements, enabling virtual characters to provide live storytelling services. Additionally, the system allows for bi-directional interaction with players through tablet touch and voice. The system aims to promote multimodal learning channels for the masses. To gain deeper insights into the effectiveness of virtual role-playing in language learning, we organized a three-week workshop to investigate our system’s impact on user experience. We extended invitations to 17 night-school freshmen from the NFU Department of Applied Foreign Languages, enrolled in an English class focused on tour guides, to participate in a virtual performance. The experiment results demonstrated that the VAILSS positively affected students’ learning outcomes, particularly in enhancing foreign language acquisition.\u0000 \u0000","PeriodicalId":442331,"journal":{"name":"網際網路技術學刊","volume":"63 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139638329","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}
《Due to the rapid development of interactive technology, sound interaction, as one of the core forms of interactive art, plays an important role in sound art. Its interactivity is one of the main factors affecting audiences’ perception. As far as the interaction of sound art is concerned, it can neither be extended as sound art nor give full play to its function, which is the key difficulty in the current development. Therefore, only by making full use of the existing advantages of virtual reality and the innovation of digital processing technology that the value of art therapy can be better reflected. However, in sound art therapy, the most critical thing is to establish the correct understanding of it and followed by the selection of the most appropriate treatment method to achieve the best outcomes. On this basis, this study reviews the progress and trend of virtual reality and related technologies. The results show that virtual reality can be used as an effective means to assist sound art therapy. Also, this paper explores how to use existing Internet technology to enhance the experience of virtual reality combined with sound art to improve the feasibility and effectiveness of treating anxiety disorders. After a systematic discussion through MFSC extraction algorithm, the results indicated that the extraction rate of MFSC changed several times and the index decreased significantly before and after the LF/HF experiment, which proved that this approach is more effective than the traditional one. In addition, VR can combine existing Internet technologies, such as high-speed Internet, cloud computing, Internet of Things technology, artificial intelligence, real-time performance, etc. Through these combined methods, the practice of the use of sound art in treating anxiety disorders will be enhanced by the interaction of virtual reality and high-speed Internet technology, which will deepen patients’ experience and improve the treatment effect.《
{"title":"Examination of the Use of VR Combined with Internet Technology to Enhance the Experience of Sound Art in the Treatment of Anxiety Disorders","authors":"Xizhi Zhang Xizhi Zhang, Huan Ding Xizhi Zhang, YuXi Xie Huan Ding","doi":"10.53106/160792642024012501002","DOIUrl":"https://doi.org/10.53106/160792642024012501002","url":null,"abstract":"\u0000 《Due to the rapid development of interactive technology, sound interaction, as one of the core forms of interactive art, plays an important role in sound art. Its interactivity is one of the main factors affecting audiences’ perception. As far as the interaction of sound art is concerned, it can neither be extended as sound art nor give full play to its function, which is the key difficulty in the current development. Therefore, only by making full use of the existing advantages of virtual reality and the innovation of digital processing technology that the value of art therapy can be better reflected. However, in sound art therapy, the most critical thing is to establish the correct understanding of it and followed by the selection of the most appropriate treatment method to achieve the best outcomes. On this basis, this study reviews the progress and trend of virtual reality and related technologies. The results show that virtual reality can be used as an effective means to assist sound art therapy. Also, this paper explores how to use existing Internet technology to enhance the experience of virtual reality combined with sound art to improve the feasibility and effectiveness of treating anxiety disorders. After a systematic discussion through MFSC extraction algorithm, the results indicated that the extraction rate of MFSC changed several times and the index decreased significantly before and after the LF/HF experiment, which proved that this approach is more effective than the traditional one. In addition, VR can combine existing Internet technologies, such as high-speed Internet, cloud computing, Internet of Things technology, artificial intelligence, real-time performance, etc. Through these combined methods, the practice of the use of sound art in treating anxiety disorders will be enhanced by the interaction of virtual reality and high-speed Internet technology, which will deepen patients’ experience and improve the treatment effect.《\u0000","PeriodicalId":442331,"journal":{"name":"網際網路技術學刊","volume":"14 2","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139638528","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 : 2024-01-01DOI: 10.53106/160792642024012501005
Manhai Li Manhai Li, Xunyi Zheng Manhai Li
Due to the impact of COVID-19 pandemic, there are quite a few concerns on organizations or companies with the capability in quickly and timely detecting psychological vulnerability level of employees and predicting potential risk behaviors. Regarding the application of daily health monitoring, ordinary camera is used to non-contact capture the shallow blood changes under the face, using algorithm to identify the physiological signs of human body, such as blood oxygen, blood pressure, respiration, and heart rate. With the rapid advancement of Internet technology, it not only assists the corresponding model of physiological signs and psychological symptoms on mapping the change regularities and health score of people’s psychological state at any time, but also utilizes suitable Internet technologies to realize and provide early warning and intervention too. The experimental results of this study show that the constant non-contact detection of human psycho-logical vulnerability level is theoretically feasible and that it can provide practical and effective reference for psychological health assessment. As a result, the scheme of epidemic psychological prevention adopted in this research has the convenience of rapid detection and real-time prediction as well as the scientific of long-time big data tracking and backtracking through the application of Internet technologies, which is a supplement to the traditional psychological detection and prevention.
{"title":"Development and Evaluation on The Framework Design of A Continuous Mental Stress Monitor Based on Contactless Sensors and Internet Technology Applications","authors":"Manhai Li Manhai Li, Xunyi Zheng Manhai Li","doi":"10.53106/160792642024012501005","DOIUrl":"https://doi.org/10.53106/160792642024012501005","url":null,"abstract":"\u0000 Due to the impact of COVID-19 pandemic, there are quite a few concerns on organizations or companies with the capability in quickly and timely detecting psychological vulnerability level of employees and predicting potential risk behaviors. Regarding the application of daily health monitoring, ordinary camera is used to non-contact capture the shallow blood changes under the face, using algorithm to identify the physiological signs of human body, such as blood oxygen, blood pressure, respiration, and heart rate. With the rapid advancement of Internet technology, it not only assists the corresponding model of physiological signs and psychological symptoms on mapping the change regularities and health score of people’s psychological state at any time, but also utilizes suitable Internet technologies to realize and provide early warning and intervention too. The experimental results of this study show that the constant non-contact detection of human psycho-logical vulnerability level is theoretically feasible and that it can provide practical and effective reference for psychological health assessment. As a result, the scheme of epidemic psychological prevention adopted in this research has the convenience of rapid detection and real-time prediction as well as the scientific of long-time big data tracking and backtracking through the application of Internet technologies, which is a supplement to the traditional psychological detection and prevention.\u0000 \u0000","PeriodicalId":442331,"journal":{"name":"網際網路技術學刊","volume":"317 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139632333","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 : 2024-01-01DOI: 10.53106/160792642024012501013
Yan Zhang Yan Zhang, Wei Wei Yan Zhang
In order to realize intelligent image recognition of foul behavior in basketball games, this paper designs a feature capture method of video foul behavior based on improved bilateral filtering algorithm. Adaptive bilateral filtering is used to denoise the video image, and the optical flow feature and HOG feature of the behavior in the denoised image are obtained by image combination feature extraction method, which is fused to a combined eigenvector. The combined features were taken as the target recognition samples, and the multi-back propagation neural network was used to identify the foul behavior features. The particle filter was used to capture the video features and identify the location of the video behavior features. The experimental results show that this method can accurately capture the changes of video behavior characteristics, and has applicable performance in the identification of foul behavior in basketball matches.
{"title":"Research on the Application of Behavioral Image Feature Capture in Basketball Game Video","authors":"Yan Zhang Yan Zhang, Wei Wei Yan Zhang","doi":"10.53106/160792642024012501013","DOIUrl":"https://doi.org/10.53106/160792642024012501013","url":null,"abstract":"\u0000 In order to realize intelligent image recognition of foul behavior in basketball games, this paper designs a feature capture method of video foul behavior based on improved bilateral filtering algorithm. Adaptive bilateral filtering is used to denoise the video image, and the optical flow feature and HOG feature of the behavior in the denoised image are obtained by image combination feature extraction method, which is fused to a combined eigenvector. The combined features were taken as the target recognition samples, and the multi-back propagation neural network was used to identify the foul behavior features. The particle filter was used to capture the video features and identify the location of the video behavior features. The experimental results show that this method can accurately capture the changes of video behavior characteristics, and has applicable performance in the identification of foul behavior in basketball matches.\u0000 \u0000","PeriodicalId":442331,"journal":{"name":"網際網路技術學刊","volume":"31 s1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139638408","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 : 2024-01-01DOI: 10.53106/160792642024012501010
Javier Carrillo-Mondejar Javier Carrillo-Mondejar, José Roldán-Gómez Javier Carrillo-Mondejar, Juan Manuel Castelo Gómez José Roldán-Gómez, Sergio Ruiz Villafranca Juan Manuel Castelo Gómez, Guillermo Suarez-Tangil Sergio Ruiz Villafranca
Since the inception of the Internet of Things (IoT), the security measures implemented on its devices have been too weak to ensure the appropriate protection of the data that they handle. Appealed by this, cybercriminals continuously seek out for vulnerable units to control, leading to attacks spreading through networks and infecting a high number of devices. On top of that, while the IoT has evolved to provide a higher degree of security, the techniques used by attackers have done so as well, which has led to the need of continuously studying the way in which these attacks are performed to gather significant knowledge for the development of the pertinent security measures. In view of this, we analyze the state of IoT attacks by developing a high-interaction honeypot for SSH and Telnet services that simulates a custom device with the ARM architecture. This study is carried out in two steps. Firstly, we analyze and classify the interaction between the attacker and the devices by clustering the commands that they sent in the compromised Telnet and SSH sessions. Secondly, we study the malware samples that are downloaded and executed in each session and classify them based on the sequence of system calls that they execute at runtime. In addition, apart from studying the active data generated by the attacker, we extract the information that is left behind after a connection with the honeypot by inspecting the metadata associated with it. In total, more than 1,578 malicious samples were collected after 9,926 unique IP addresses interacted with the system, with the most downloaded malware family being Hajime, with 70.5% of samples belonging to it, and also detecting others such as Mirai, Gafgyt, Dofloo and Xorddos.
{"title":"Stories from a Customized Honeypot for the IoT","authors":"Javier Carrillo-Mondejar Javier Carrillo-Mondejar, José Roldán-Gómez Javier Carrillo-Mondejar, Juan Manuel Castelo Gómez José Roldán-Gómez, Sergio Ruiz Villafranca Juan Manuel Castelo Gómez, Guillermo Suarez-Tangil Sergio Ruiz Villafranca","doi":"10.53106/160792642024012501010","DOIUrl":"https://doi.org/10.53106/160792642024012501010","url":null,"abstract":"\u0000 Since the inception of the Internet of Things (IoT), the security measures implemented on its devices have been too weak to ensure the appropriate protection of the data that they handle. Appealed by this, cybercriminals continuously seek out for vulnerable units to control, leading to attacks spreading through networks and infecting a high number of devices. On top of that, while the IoT has evolved to provide a higher degree of security, the techniques used by attackers have done so as well, which has led to the need of continuously studying the way in which these attacks are performed to gather significant knowledge for the development of the pertinent security measures. In view of this, we analyze the state of IoT attacks by developing a high-interaction honeypot for SSH and Telnet services that simulates a custom device with the ARM architecture. This study is carried out in two steps. Firstly, we analyze and classify the interaction between the attacker and the devices by clustering the commands that they sent in the compromised Telnet and SSH sessions. Secondly, we study the malware samples that are downloaded and executed in each session and classify them based on the sequence of system calls that they execute at runtime. In addition, apart from studying the active data generated by the attacker, we extract the information that is left behind after a connection with the honeypot by inspecting the metadata associated with it. In total, more than 1,578 malicious samples were collected after 9,926 unique IP addresses interacted with the system, with the most downloaded malware family being Hajime, with 70.5% of samples belonging to it, and also detecting others such as Mirai, Gafgyt, Dofloo and Xorddos.\u0000 \u0000","PeriodicalId":442331,"journal":{"name":"網際網路技術學刊","volume":"66 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139632891","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 : 2024-01-01DOI: 10.53106/160792642024012501006
Shu-Chuan Chu Shu-Chuan Chu, Ting-Ting Wang Shu-Chuan Chu, Ali Riza Yildiz Ting-Ting Wang, Jeng-Shyang Pan Ali Riza Yildiz
In recent years, more and more problems in the industry have started to be solved using metaheuristics. In this paper, inspired by the ship maneuvering motion function and the rescue process, we propose ship rescue optimization (SRO) to solve the challenging optimization problem. The ship rescue process can be divided into two types of delayed rescue (large area rescue) and immediate rescue (small rescue) according to the searched person, and we can correspond these two types of rescue behaviors to the search space exploration and exploitation processes, respectively. In this process, SRO simulates the motion process of ship rescue according to the ship maneuvering equation of motion and finally comes up with an optimal position update algorithm. We verified the performance of the proposed algorithm on different dimensions of 57 test functions consisting of CEC2013 and CEC2017 and on three real engineering problems, and compared it with eight current mainstream algorithms. The algorithm proposed in this paper is shown to be robustly applicable in solving challenging optimization problems.
{"title":"Ship Rescue Optimization: A New Metaheuristic Algorithm for Solving Engineering Problems","authors":"Shu-Chuan Chu Shu-Chuan Chu, Ting-Ting Wang Shu-Chuan Chu, Ali Riza Yildiz Ting-Ting Wang, Jeng-Shyang Pan Ali Riza Yildiz","doi":"10.53106/160792642024012501006","DOIUrl":"https://doi.org/10.53106/160792642024012501006","url":null,"abstract":"\u0000 In recent years, more and more problems in the industry have started to be solved using metaheuristics. In this paper, inspired by the ship maneuvering motion function and the rescue process, we propose ship rescue optimization (SRO) to solve the challenging optimization problem. The ship rescue process can be divided into two types of delayed rescue (large area rescue) and immediate rescue (small rescue) according to the searched person, and we can correspond these two types of rescue behaviors to the search space exploration and exploitation processes, respectively. In this process, SRO simulates the motion process of ship rescue according to the ship maneuvering equation of motion and finally comes up with an optimal position update algorithm. We verified the performance of the proposed algorithm on different dimensions of 57 test functions consisting of CEC2013 and CEC2017 and on three real engineering problems, and compared it with eight current mainstream algorithms. The algorithm proposed in this paper is shown to be robustly applicable in solving challenging optimization problems.\u0000 \u0000","PeriodicalId":442331,"journal":{"name":"網際網路技術學刊","volume":"10 11","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139637725","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 : 2024-01-01DOI: 10.53106/160792642024012501009
Zhongxiang Zhan Zhongxiang Zhan, Sai Ji Zhongxiang Zhan, Wenying Zheng Sai Ji, Dengzhi Liu Wenying Zheng
Android is an open-source mobile operating system, with more than 70% of the mobile market share, widely popular on various intelligent devices. At the same time, the number of new malicious applications keeps increasing every year. In this paper, we first discuss the advantages and disadvantages of various detection methods for malicious software. A single detection method can only cover specific types of malware. Therefore, we propose a system that combines static structural analysis and dynamic detection of malware. This system has dual detection capability, which consists of a client and a server. The client is a lightweight Android application that is used to obtain the relevant data information of the installation package. The server is responsible for static analysis of APK and dynamic running of monitoring logs to get the relevant feature information. Based on the feature information, the Bagging algorithm of ensemble learning is adopted, and the decision tree and random forest are combined to identify the malware accurately. We collected 4210 Android software samples, with malicious apps accounting for about 20% of the total. Cross-testing of malware detection on this sample set showed that DroidExaminer achieved approximately 96% accuracy in detecting malware. It can resist confusion and conversion techniques, and the test performance overhead is less. In addition, DroidExaminer can alert the user to the details of malware intrusion so that the user can prevent malware intrusion.
{"title":"DroidExaminer: An Android Malware Hybrid Detection System Based on Ensemble Learning","authors":"Zhongxiang Zhan Zhongxiang Zhan, Sai Ji Zhongxiang Zhan, Wenying Zheng Sai Ji, Dengzhi Liu Wenying Zheng","doi":"10.53106/160792642024012501009","DOIUrl":"https://doi.org/10.53106/160792642024012501009","url":null,"abstract":"\u0000 Android is an open-source mobile operating system, with more than 70% of the mobile market share, widely popular on various intelligent devices. At the same time, the number of new malicious applications keeps increasing every year. In this paper, we first discuss the advantages and disadvantages of various detection methods for malicious software. A single detection method can only cover specific types of malware. Therefore, we propose a system that combines static structural analysis and dynamic detection of malware. This system has dual detection capability, which consists of a client and a server. The client is a lightweight Android application that is used to obtain the relevant data information of the installation package. The server is responsible for static analysis of APK and dynamic running of monitoring logs to get the relevant feature information. Based on the feature information, the Bagging algorithm of ensemble learning is adopted, and the decision tree and random forest are combined to identify the malware accurately. We collected 4210 Android software samples, with malicious apps accounting for about 20% of the total. Cross-testing of malware detection on this sample set showed that DroidExaminer achieved approximately 96% accuracy in detecting malware. It can resist confusion and conversion techniques, and the test performance overhead is less. In addition, DroidExaminer can alert the user to the details of malware intrusion so that the user can prevent malware intrusion.\u0000 \u0000","PeriodicalId":442331,"journal":{"name":"網際網路技術學刊","volume":"25 30","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139631736","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 : 2024-01-01DOI: 10.53106/160792642024012501007
Zhichao Yang Zhichao Yang, Yan Kong Zhichao Yang, Chih-Hsien Hsia Yan Kong
In recent years, with the increasingly severe traffic environment, most cities are facing various traffic congestion problems, and the demand for intelligent regulation of traffic signals is also increasing. In this study, we propose a new intelligent traffic light control algorithm, dual experience replay light (DERLight), which innovatively and efficiently designs a dual experience replay training mechanism based on the classic deep Q network (DQN) framework and considers the dynamic epoch function. As results show that compared with some state-of-the-art algorithms, DERLight can shorten the average travel time of vehicles, increase the throughput at intersections, and also speed up the convergence of the network. In addition, the design of this algorithm framework is not only limited to the field of intelligent transportation, but also has transferability for some other fields.
{"title":"DERLight: A Deep Reinforcement Learning Traffic Light Control Algorithm with Dual Experience Replay","authors":"Zhichao Yang Zhichao Yang, Yan Kong Zhichao Yang, Chih-Hsien Hsia Yan Kong","doi":"10.53106/160792642024012501007","DOIUrl":"https://doi.org/10.53106/160792642024012501007","url":null,"abstract":"\u0000 In recent years, with the increasingly severe traffic environment, most cities are facing various traffic congestion problems, and the demand for intelligent regulation of traffic signals is also increasing. In this study, we propose a new intelligent traffic light control algorithm, dual experience replay light (DERLight), which innovatively and efficiently designs a dual experience replay training mechanism based on the classic deep Q network (DQN) framework and considers the dynamic epoch function. As results show that compared with some state-of-the-art algorithms, DERLight can shorten the average travel time of vehicles, increase the throughput at intersections, and also speed up the convergence of the network. In addition, the design of this algorithm framework is not only limited to the field of intelligent transportation, but also has transferability for some other fields.\u0000 \u0000","PeriodicalId":442331,"journal":{"name":"網際網路技術學刊","volume":"52 s39","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139633829","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 recent years, OCR data has been used for learning and analyzing document classification. In addition, some neural networks have used image recognition for training, such as the network published by the ImageNet Large Scale Visual Recognition Challenge for document image training, AlexNet, GoogleNet, and MobileNet. Document image classification is important in data extraction processes and often requires significant computing power. Furthermore, it is difficult to implement image classification using general computers without a graphics processing unit (GPU). Therefore, this study proposes a lightweight neural network application that can perform document image classification on general computers or the Internet of Things (IoT) without a GPU. Plustek Inc. provided 3065 receipts belonging to 58 categories. Three datasets were considered as test samples while the remaining were considered as training samples to train the network to obtain a classifier. After the experiments, the classifier achieved 98.26% accuracy, and only 3 out of 174 samples showed errors.
{"title":"Document Classification Using Lightweight Neural Network","authors":"Chung-Hsing Chen Chung-Hsing Chen, Ko-Wei Huang Chung-Hsing Chen","doi":"10.53106/160792642023122407012","DOIUrl":"https://doi.org/10.53106/160792642023122407012","url":null,"abstract":"In recent years, OCR data has been used for learning and analyzing document classification. In addition, some neural networks have used image recognition for training, such as the network published by the ImageNet Large Scale Visual Recognition Challenge for document image training, AlexNet, GoogleNet, and MobileNet. Document image classification is important in data extraction processes and often requires significant computing power. Furthermore, it is difficult to implement image classification using general computers without a graphics processing unit (GPU). Therefore, this study proposes a lightweight neural network application that can perform document image classification on general computers or the Internet of Things (IoT) without a GPU. Plustek Inc. provided 3065 receipts belonging to 58 categories. Three datasets were considered as test samples while the remaining were considered as training samples to train the network to obtain a classifier. After the experiments, the classifier achieved 98.26% accuracy, and only 3 out of 174 samples showed errors.","PeriodicalId":442331,"journal":{"name":"網際網路技術學刊","volume":"17 2","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139188001","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 : 2023-12-01DOI: 10.53106/160792642023122407001
Changqing Zhao Changqing Zhao, Ling Xia Liao Changqing Zhao, Han-Chieh Chao Ling Xia Liao, Roy Xiaorong Lai Han-Chieh Chao, Miao Zhang Roy Xiaorong Lai
While Software-Defined Networks (SDNs) have separated control and data planes and completely decouple the flow control from the data forwarding to enable network flexibility, programmability, and innovation, they also raise serious security concerns in each plane and the interfaces between the two planes. This paper, instead of studying the security issues in the SDN control plane as many literatures have done in current research, focuses on the security issues in the SDN data plane, aiming at the state of the art mechanims to identify, detect, and mitigate them. Specifically, this paper reviews the typical models, detections, and mitigations of SDN flow table overflow attacks. After reviewing the various vulnerabilities in SDNs, this paper categorizes the flow table overflow attacks into saturation, low-rate table exhaustion, and slow saturation attacks, and summarizes the attack models, detections, and mitigations of each category. It reviews the typical attacks that can overflow the flow tables and provides the main challenges and open issues for the future research.
软件定义网络（SDN）将控制平面和数据平面分离，并将流量控制与数据转发完全解耦，从而实现了网络的灵活性、可编程性和创新性，但同时也在每个平面以及两个平面之间的接口上引发了严重的安全问题。本文并不像当前研究中的许多文献那样研究 SDN 控制平面的安全问题，而是重点关注 SDN 数据平面的安全问题，旨在研究识别、检测和缓解这些问题的最新机制。具体来说，本文回顾了 SDN 流量表溢出攻击的典型模型、检测和缓解方法。在回顾了 SDN 中的各种漏洞后，本文将流表溢出攻击分为饱和攻击、低速率表耗尽攻击和慢速饱和攻击，并总结了每类攻击的攻击模型、检测和缓解方法。本文回顾了可能导致流量表溢出的典型攻击，并提出了未来研究的主要挑战和开放性问题。
{"title":"Flow Table Overflow Attacks in Software Defined Networks: A Survey","authors":"Changqing Zhao Changqing Zhao, Ling Xia Liao Changqing Zhao, Han-Chieh Chao Ling Xia Liao, Roy Xiaorong Lai Han-Chieh Chao, Miao Zhang Roy Xiaorong Lai","doi":"10.53106/160792642023122407001","DOIUrl":"https://doi.org/10.53106/160792642023122407001","url":null,"abstract":"While Software-Defined Networks (SDNs) have separated control and data planes and completely decouple the flow control from the data forwarding to enable network flexibility, programmability, and innovation, they also raise serious security concerns in each plane and the interfaces between the two planes. This paper, instead of studying the security issues in the SDN control plane as many literatures have done in current research, focuses on the security issues in the SDN data plane, aiming at the state of the art mechanims to identify, detect, and mitigate them. Specifically, this paper reviews the typical models, detections, and mitigations of SDN flow table overflow attacks. After reviewing the various vulnerabilities in SDNs, this paper categorizes the flow table overflow attacks into saturation, low-rate table exhaustion, and slow saturation attacks, and summarizes the attack models, detections, and mitigations of each category. It reviews the typical attacks that can overflow the flow tables and provides the main challenges and open issues for the future research.","PeriodicalId":442331,"journal":{"name":"網際網路技術學刊","volume":"5 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139188221","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：