作者机构:
[Zhou, Jiaqi; Wu, Wenfeng; Yang, Xin; Jiang, Lingfeng] College of Electrical and Information Engineering, Changsha University of Science and Technology, Changsha;410114, China;[Zhou, Jiaqi; Wu, Wenfeng; Yang, Xin; Jiang, Lingfeng] 410114, China
会议名称:
10th China International Conference on Electricity Distribution, CICED 2022
作者机构:
[Ding, Xiangling] Hunan Univ Sci & Technol, Xiangtan, Peoples R China.;[Ding, Xiangling] Shenzhen Key Lab Media Secur, Guangdong Key Lab Intelligent Informat Proc, Shenzhen, Peoples R China.;[Huang, Pu; Zhang, Dengyong] Changsha Univ Sci & Technol, Changsha, Peoples R China.;[Zhao, Xianfeng; Ding, Xiangling] Chinese Acad Sci, State Key Lab Informat Secur, Beijing, Peoples R China.;[Zhao, Xianfeng] Univ Chinese Acad Sci, Sch Cyber Secur, Beijing, Peoples R China.
会议名称:
47th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
会议时间:
MAY 22-27, 2022
会议地点:
Singapore, SINGAPORE
会议主办单位:
[Ding, Xiangling] Hunan Univ Sci & Technol, Xiangtan, Peoples R China.^[Ding, Xiangling] Shenzhen Key Lab Media Secur, Guangdong Key Lab Intelligent Informat Proc, Shenzhen, Peoples R China.^[Huang, Pu;Zhang, Dengyong] Changsha Univ Sci & Technol, Changsha, Peoples R China.^[Ding, Xiangling;Zhao, Xianfeng] Chinese Acad Sci, State Key Lab Informat Secur, Beijing, Peoples R China.^[Zhao, Xianfeng] Univ Chinese Acad Sci, Sch Cyber Secur, Beijing, Peoples R China.^[Ding, Xiangling] Guangdong Prov Key Lab Informat Secur Technol, Guangzhou, Peoples R China.
会议论文集名称:
International Conference on Acoustics Speech and Signal Processing ICASSP
关键词:
Video frame interpolation;bidirectional encoder;channel attention cascade;local lightweight
摘要:
Deep Neural Networks based video frame interpolation, synthesizing in-between frames given two consecutive neighboring frames, typically depends on heavy model architectures, preventing them from being deployed on small terminals. When directly adopting the lightweight network architecture from these models, the synthesized frames may suffer from poor visual appearance. In this paper, a lightweight-driven video frame interpolation network (L-2 BEC2) is proposed. Concretely, we first improve the visual appearance by introducing the bidirectional encoding structure with channel attention cascade to better characterize the motion information; then we further adopt the local network lightweight idea into the aforementioned structure to significantly eliminate its redundant parts of the model parameters. As a result, our L-2 BEC2 performs favorably at the cost of only one third of the parameters compared with the state-of-the-art methods on public datasets. Our source code is available at https : //github.com/Pumpkin123709/LBEC.git.
摘要:
A defect detection method based on an improved Faster R-CNN is proposed in this paper, which deals with the problem of poor detection accuracy of small target defects in the surface defect detection of hot rolled steel plates. The feature extraction network of Faster R-CNN is modified for the defect features on the surface of steel plates to obtain better defect detection results. The experimental results show that the improved network has a good improvement in accuracy for small defect detection.
作者机构:
[He, Yingqian; Zhou, Lixing] Changsha University of Science & Technology, School of Electrical and Information Engineering, Changsha, China;[Chen, Yaohong] State Grid Hunan Electric Power Company Limited, Changsha, China
会议名称:
10th China International Conference on Electricity Distribution, CICED 2022
摘要:
Compared with cloud computing, Mobile Edge Computing (MEC) can sink some services and functions located in cloud servers to edge nodes to reduce network latency and provide real-time services. However, MEC not only inherits the security issues in cloud computing, but also faces new security risks. To ensure the security and privacy of messages transmitted in the channel, a secure Authentication and Key Agreement (AKA) protocol is essential. However, the existing AKA protocols are not lightweight enough or require a cloud server or a trusted third party to participate in the authentication process. Therefore, this paper designs a fast AKA protocol based on time-sensitive token for MEC. With this protocol, the terminal node can achieve fast authentication through the applied token, and the authentication process only requires a related edge node to participate. Simulation results based on ProVerif and informal security analysis show that our protocol can resist various common attacks. The comparison with related protocols shows that our protocol only needs to spend very little computational and communicational cost to authenticate a terminal node with a token.
作者机构:
[Li, Juntang] State Grid Hunan Extra High Voltage Substation Company, Substation Intelligent Operation and Inspection Laboratory, State Grid Hunan Electric Power Co., Ltd, Changsha, China;[Ouyang, Juan] Hunan Communication Polytechnic College, Changsha, China;[Huang, Yifei] College of Electrical and Information Engineering, Changsha University of Science and Technology, Changsha, China;[Hu, Wentao] Hunan Electric Power Design Institute Co., Ltd., China Energy Engineering Group, Changsha, China
会议名称:
10th China International Conference on Electricity Distribution, CICED 2022
摘要:
Convolutional Neural Network (CNN) is more and more widely used in various fileds, and its computation and memory-demand are also increasing significantly. In order to make it applicable to limited conditions such as embedded application, network compression comes out. Among them, researchers pay more attention to network pruning. In this paper, we encode the convolution network to obtain the similarity of different encoding nodes, and evaluate the connectivity-power among convolutional kernels on the basis of the similarity. Then impose different level of penalty according to different connectivity-power. Meanwhile, we propose Channel Pruning base on the Dissimilarity of Angle (DACP). Firstly, we train a sparse model by GL penalty, and impose an angle dissimilarity (AD) constraint on the channels and filters of convolutional network to obtain a more sparse structure. Eventually, the effectiveness of our method is demonstrated in the section of experiment. On CIFAR-10, we reduce 66.86% FLOPs on VGG-16 with 93.31% accuracy after pruning, where FLOPs represents the number of floating-point operations per second of the model. Moreover, on ResNet-32, we reduce FLOPs by 58.46%, which makes the accuracy after pruning reach 91.76%.
作者机构:
[Zhou, Xiu; Tian, Tian; Bai, Jin; Luo, Yan] Power Research Institute of State, Grid State Grid Ningxia Power Co., Ltd, Ningxia;750002, China;[Yang, Xin; Wu, Wenfeng] College of Electrical and Information Engineering, Changsha University of Science and Technology, Changsha;410114, China;[Zhou, Xiu; Tian, Tian; Bai, Jin; Luo, Yan] 750002, China
会议名称:
10th China International Conference on Electricity Distribution, CICED 2022
作者机构:
[Lin, Songchen; Li, Jinyan] Power Planning Center of Hainan Electric Power Design, Research Institute of PowerChina Co., Ltd, Hainan, Haikou, China;[Liu, Keming; Huang, Jingjie; Sun, Chenhao; Qin, Zhichao] Smart Grids Operation and Control Key Laboratory of Hunan Province (CSUST), Changsha University of Science and Technology, Hunan, Changsha, China
会议名称:
10th China International Conference on Electricity Distribution, CICED 2022
作者机构:
[Liang, Zhao; Liang, Xiaolin; Chen, Libo] Changsha Univ Sci & Technol, Sch Math & Stat, Changsha, Peoples R China.;[Huang, Yajuan] Chinese Acad Sci, Inst Informat Engn, Changsha Univ Sci & Technol, Changsha, Peoples R China.;[Zhu, He; Liu, Wen] Chinese Acad Sci, Inst Informat Engn, Beijing, Peoples R China.
会议名称:
IEEE International Conference on Fuzzy Systems (FUZZ-IEEE) / IEEE World Congress on Computational Intelligence (IEEE WCCI) / International Joint Conference on Neural Networks (IJCNN) / IEEE Congress on Evolutionary Computation (IEEE CEC)
会议时间:
JUL 18-23, 2022
会议地点:
Padua, ITALY
会议主办单位:
[Liang, Xiaolin;Liang, Zhao;Chen, Libo] Changsha Univ Sci & Technol, Sch Math & Stat, Changsha, Peoples R China.^[Huang, Yajuan] Chinese Acad Sci, Inst Informat Engn, Changsha Univ Sci & Technol, Changsha, Peoples R China.^[Liu, Wen;Zhu, He] Chinese Acad Sci, Inst Informat Engn, Beijing, Peoples R China.
会议论文集名称:
IEEE International Joint Conference on Neural Networks (IJCNN)
关键词:
Hate Meme Detection;Triplet-Relation;Cross-Attention;Image Caption
摘要:
Memes are spreading on social networking. Most are created to be humorous, while some become hateful with the combination of images and words, conveying negative information to people. The hateful memes detection poses an interesting multi-modal fusion problem, unlike traditional multi-modal tasks, the majority of memos have images and text that are only weakly consistent or even uncorrelated, so various modalities contained in the data play an important role in predicting its results. In this paper, we attempt to work on the Facebook Meme challenge, which solves the binary classification task of predicting a meme's hatefulness or not. We extract triplet-relation information from origin OCR text features, image content features and image caption features and proposed a novel cross-attention network to address this task. TRICAN leverages object detection and image caption models to explore visual modalities to obtain "actual captions" and then combines combine origin OCR text with the multi-modal representation to perform hateful memes detection. These meme-related features are then reconstructed and fused into one feature vector for prediction. We have performed extensively experimental on multi-modal memory datasets. Experimental results demonstrate the effectiveness of TRICAN and the usefulness of triplet-relation information.
会议名称:
7th IEEE International Conference on Advanced Robotics and Mechatronics
会议时间:
JUL 09-11, 2022
会议地点:
Guilin, PEOPLES R CHINA
会议主办单位:
[Xu, Tao;Yuan, Xiaofang;Zhong, Hang] Hunan Univ, Coll Elect & Informat Engn, Changsha, Hunan, Peoples R China.^[Zhang, Hui] Hunan Univ, Natl Engn Res Ctr, Coll Robot, Changsha, Hunan, Peoples R China.^[Zhang, Hui] Hunan Univ, Natl Engn Res Ctr, Robot Visual Percept & Control Technol, Changsha, Hunan, Peoples R China.^[Zhou, Xidong;Tan, Xuan;Zhang, Jieqingxin] Changsha Univ Sci & Technol, Coll Elect & Informat Engn, Changsha, Hunan, Peoples R China.
摘要:
At present, the indoor positioning methods applied to robots are still subject to many limitations, which restrict the further development of indoor robots. Among the existing localization methods, ultra-wideband (UWB) technology can realize low-cost centimeter-level localization, but the localization accuracy is easily affected by Non-Line-Of-Sight (NLOS). However, using an odometry for position measurement is less susceptible to external environment interference, but it has accumulated errors. In this paper, a Weighted Adaptive Kalman Filter (WAKF) positioning method based on UWB and odometry is proposed, which eliminates the influence of the cumulative error of the odometry on the positioning accuracy to a certain extent, and solves the problem of NLOS positioning error caused by indoor obstacles, to realize the positioning of the robot in complex indoor scenes. The method uses the Kalman Filtering(KF) algorithm to fuse UWB data and odometry data, uses the power difference to identify Line-Of-Sight(LOS) and NLOS scenes, and adaptively adjusts the Kalman filter weights. The method is tested on the Epidemic Prevention and Disinfection Robot (EPADR). The experimental results show that this method meets the demand of indoor positioning, can eliminate the influence of the cumulative error of the odometry and effectively improve the positioning accuracy of the robot in the NLOS scene. Its average positioning error is 0.16m. This method improves the positioning accuracy and robustness of the robot in the indoor environment.
会议论文集名称:
IEEE International Conference on Trust Security and Privacy in Computing and Communications
关键词:
Platoon;False data injection attack;Sequential importance sampling;Intrusion detection,Threshold
摘要:
Intelligent connected vehicle platoon technology can reduce traffic congestion and vehicle fuel. However, attacks on the data transmitted by the platoon are one of the primary challenges encountered by the platoon during its travels. The false data injection (FDI) attack can lead to road congestion and even vehicle collisions, which can impact the platoon. However, the complexity of the cellular - vehicle to everything (C-V2X) environment, the single source of the message and the poor data processing capability of the on board unit (OBU) make the traditional detection methods' success rate and response time poor. This study proposes a platoon state information fusion method using the communication characteristics of the platoon in C-V2X and proposes a novel platoon intrusion detection model based on this fusion method combined with sequential importance sampling (SIS). The SIS is a measured strategy of Monte Carlo integration sampling. Specifically, the method takes the status information of the platoon members as the predicted value input. It uses the leader vehicle status information as the posterior probability of the observed value to the current moment of the platoon members. The posterior probabilities of the platoon members and the weights of the platoon members at the last moment are used as input to update the weights of the platoon members at the current moment and obtain the desired platoon status information at the present moment. Moreover, it compares the status information of the platoon members with the desired status information to detect attacks on the platoon. Finally, the effectiveness of the method is demonstrated by simulation.
作者机构:
[Luo, Dingyuan] School of Electrical Engineering, Xi'an Jiaotong University, Xi'an, China;[Chen, Zimin] School of Electrical & Information Engineering, Changsha University of Science and Technology, Changsha, China;[Yu, Jiaqi] College of Electronic Information and Electrical Engineering, Changsha University, Changsha, China;[Xu, Yong] Research and Development Center, State Grid Hunan Integrated Energy Service Co., Ltd., Changsha, China
会议名称:
10th China International Conference on Electricity Distribution, CICED 2022
会议论文集名称:
IEEE International Symposium on Power Electronics for Distributed Generation Systems
关键词:
Cascaded H-bridge PV system;quadruple active bridges;reduced capacitance;low-frequency voltage ripple;fluctuation ratio;reliability
摘要:
Cascaded H-bridge (CHB) multilevel converters are promising candidates for large-scale grid-connected PV systems thanks to its modularity, scalability and distributed maximum power point tracking (DMPPT). However, CHB converters inherently have low-frequency power ripple in dc links, thus require large dc-link electrolytic capacitors to mitigate the resulting voltage ripple, which will lead in reduction of system reliability. In order to solve the above issues, this paper adopts the system topology architecture composed of CHB multilevel converter and quadruple active bridges (QAB) dc-dc converters, combining with the intercross control structure. This scheme not only can improve the system reliability by decreasing the dc-link capacitance, but also can prevent the low-frequency ripple propagating to the PV ports. Simulation results clearly verify the effectiveness and feasibility of the topology and control strategy.
会议名称:
20th International Workshop on Digital-Forensics and Watermarking (IWDW)
会议时间:
NOV 20-22, 2021
会议地点:
Beijing, PEOPLES R CHINA
会议主办单位:
[Zhang, Le-Bing;Cai, Juan] Huaihua Univ, Huaihua 418000, Peoples R China.^[Peng, Fei] Hunan Univ, Changsha 410082, Peoples R China.^[Long, Min] Changsha Univ Sci & Technol, Changsha 410114, Peoples R China.
会议论文集名称:
Lecture Notes in Computer Science
关键词:
Face morphing detection;Morphing artifacts;Attention convolutional neural network;Multiple scales
摘要:
Face morphing attack is becoming a serious threat to the existing face recognition systems. Some different approaches for morphing attack detection have been put forward. However, there are few methods concentrated on detecting the morphing artifacts, which often appear in morphed facial images. In this work, we propose a multiple scales attention convolutional neural network (MSA-CNN), a novel approach that can effectively detect the morphing artifacts in face morphing attacks. It utilizes the attention mechanism to continuously pay attention to the morphing artifacts in multiple scales and finally realizes face morphing detection. Experimental results and analysis show that it can effectively locate the region of morphing artifacts, and outperforms the existing deep learning-based blind face morphing detection frameworks.
作者机构:
[Yue, Yufei; Wu, Xinglong; Yang, Xi; Wang, Wen; Tang, Xin] School of Electrical and Information Engineering, Changsha University of Science & Technology, Changsha, China;[Guo, Peng] National Electric Power Conversion and Control Engineering Technology Research Center, Hunan University, Changsha, China
会议名称:
10th China International Conference on Electricity Distribution, CICED 2022
作者机构:
[Guo, Zhijie; Su, Jianhong; Liu, Yanzheng] International College of Engineering, Changsha University of Science &technology, Changsha, China;[Sun, Chenhao] School of Electrical & Information Engineering, Changsha University of Science & Technology, Changsha, China;[Liao, Jinrong] Project Management Department, Yueding Power Engineering Supervision Co., Ltd., ZhaoQing, China;[Liu, Shengxiang] Technical Service Center, State Grid Hunan Extra High Voltage, Substation Company, Changsha, China
会议名称:
10th China International Conference on Electricity Distribution, CICED 2022
作者机构:
[Xu, Zhuoran; Sun, Zhengjie; Sun, Chenhao] School of Electrical & Information Engineering, Changsha University of Science & Technology, Changsha, China;[Liu, Yanzheng] International College of Engineering, Changsha University of Science & Technology, Changsha, China;[Sun, Ruping] Shuangfeng County Power Supply Branch, State Grid of China, Loudi, China;[Liu, Shengxiang] Hunan Extra High Voltage Substation Company, State Grid of China, Changsha, China
会议名称:
10th China International Conference on Electricity Distribution, CICED 2022