检索结果-内蒙古大学图书馆

Edge Computing and Scalable Cloud (EdgeCom), IEEE International Conference on

作者： Wei Chen State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and Engineering Beihang University Beijing China

ISBN: (数字)9798350377132

ISBN: (纸本)9798350377149

Dense video captioning involves detecting and describing multiple events within a video coherently. Events within a video typically share a common topic, and incorporating this topic information into the model can enhance the quality and coherence of the generated captions. However, existing dense video captioning datasets lack explicit topic annotations. To address this, we design a topic generator that utilizes a diffusion model with strong generative capabilities to generate topic information from a given video. During the training stage, we use the features of ground-truth captions as pseudo-topic labels. The video topics diffuse from the features of groundtruth captions to a random distribution, and the model learns to reverse this noising process conditioned on video features. During inference, the model iteratively denoises the Gaussian noise into topic features conditioned on video features. In this paper, we propose TopicDVC, a novel dense video captioning framework. TopicDVC applies the topic information generated by the topic generator to guide the model in generating more coherent captions. Experiments on the ActivityNet Captions dataset demonstrate that leveraging the topics generated by the diffusion model significantly improves the performance of dense video captioning, producing more accurate and coherent captions.

关键词： Training Accuracy Annotations Video description Gaussian noise Coherence Diffusion models

来源：评论

学校读者我要写书评

暂无评论

An Advanced Approach for performing Cyber Fraud using Banner Grabbing 5

An Advanced Approach for performing Cyber Fraud using Banner...

引用

5th International Conference on Contemporary Computing and Informatics, IC3I 2022

作者： Kaushik, Keshav Punhani, Ishika Sharma, Smita Martolia, Mamta University of Petroleum and Energy Studies School of Computer Science Dehradun India Amity University Department of Electronics and Communication Engineering Uttar Pradesh India G L Bajaj Institute of Technology Management Greater Noida Department of Computer Science and Engineering Uttar Pradesh India

ISBN: (纸本)9798350398267

In this paper, the authors have discussed how cyber-attackers or hackers can do cyber fraud or can steal information from some systems using the banner grabbing approach. Cybercriminals and security personnel utilize the banner grabbing technique to gather data about a computer network or a system as well as the services operating on its open ports. An informational banner is a group of text shown by a web application that displays details like the kind and edition of software being used by a laptop or web application. The welcome screens pro-vide information about network hosts' software versions and other infrastructure components, giving attackers an edge in network assaults. In this paper, the authors have targeted the vulnerable metasploitable machine to perform the banner grabbing. Moreover, the authors have also highlighted the preventive measures against banner grabbing. The authors have employed several techniques for capturing banners, including telnet, Netcat, curl, Nmap, and Metasploit. This data may be utilized for reconnaissance and, eventually, improved exploitation. © 2022 IEEE.

关键词： Crime

来源：评论

学校读者我要写书评

暂无评论

Deep Unfolding 3D Non-Local Transformer Network for Hyperspectral Snapshot Compressive Imaging

Deep Unfolding 3D Non-Local Transformer Network for Hyperspe...

引用

IEEE International Conference on Multimedia and Expo (ICME)

作者： Zheng Zhou Zongxin Liu Yongyong Chen Bingzhi Chen Biqing Zeng Yicong Zhou School of Electronics and Communication Engineering Guangzhou University Guangzhou China School of Software South China Normal University Foshan China School of Computer Science and Technology Harbin Institute of Technology Shenzhen China Department of Computer and Information Science University of Macau Macau China

ISBN: (数字)9798350390155

ISBN: (纸本)9798350390162

Hyperspectral compressive imaging has shown remarkable advancements through the adoption of deep unfolding frameworks, which integrate the proximal mapping prior into the data fidelity term to formulate the reconstruction problem. However, existing technologies still face challenges in effectively capturing spatial-spectral features during the iterative deep prior learning stage, leading to unsatisfactory performance degradation. To address this issue, we propose a deep unfolding 3D non-local transformer (3DNLT) network for hyperspectral compressive imaging. A learnable half-quadratic splitting (HQS) algorithm is utilized to iteratively update the linear projection. Furthermore, a 3D non-local attention ushaped transformer is presented as the deep proximal mapping prior module to obtain the spatial-spectral long-range dependency features, leading to enhance the network’s ability to capture fine-grained hyperspectral and spatial details. Experimental results on both synthetic and real hyperspectral image reconstruction have demonstrated the superior performance of the 3DNLT network compared to state-of-the-art methods.

关键词： Degradation Three-dimensional displays Image coding Imaging Transformers Robustness Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Frequency Agile Strategy Design of Radar Network for Wideband Sweep Jamming via Coalition Formation Game

引用

IEEE Transactions on Aerospace and Electronic Systems 2025年第3期61卷 7906-7919页

作者： Liao, Maosen Xiong, Kui Zhang, Long Yu, Xianxiang Cui, Guolong University of Electronic Science and Technology of China School of Information and Communication Engineering Chengdu611731 China Laboratory of Electromagnetic Space Cognition and Intelligent Control Technology Beijing100089 China School of Information and Communication Engineering Chengdu611731 China UESTC Shenzhen Institute for Advanced Study Chengdu611731 China

This article proposes an intelligent frequency agile design strategy for radar network using coalition game theory to counter wideband sweep jamming, which is formulated as a two-stage decision-making problem. In the first stage, the entire frequency agile range is divided into several non-overlapping agile subrange, and each radar selects a frequency agile subrange to join, which is solved by the coalition formation algorithm. Then, after achieving a stable coalition partition, a distributed multi-agent reinforcement learning algorithm is developed to address the specific carrier frequency selection problem within each coalition. By combining these two stages of the frequency agile design strategy, a joint coalition formation game and distributed Q-learning algorithm (CFDQ) is proposed. Finally, numerical simulations are conducted to illustrate the superiority of the proposed algorithm over several typical frequency agile strategies. © 1965-2011 IEEE.

关键词： Jamming

来源：评论

学校读者我要写书评

暂无评论

Toward Realistic Camouflaged Object Detection: Benchmarks and Method

arXiv

引用

arXiv 2025年

作者： Xin, Zhimeng Wu, Tianxu Chen, Shiming Ye, Shuo Xie, Zijing Zou, Yixiong You, Xinge Guo, Yufei The School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan430074 China The School of Electronic Information and Communications Huazhong University of Science and Technology Wuhan430074 China The School of Computer Science & Technology Huazhong University of Science and Technology Wuhan430074 China The Intelligent Science and Technology Academy of CASIC Beijing100041 China

Camouflaged object detection (COD) primarily relies on semantic or instance segmentation methods. While these methods have made significant advancements in identifying the contours of camouflaged objects, they may be inefficient or cost-effective for tasks that only require the specific location of the object. Object detection algorithms offer an optimized solution for Realistic Camouflaged Object Detection (RCOD) in such cases. However, detecting camouflaged objects remains a formidable challenge due to the high degree of similarity between the features of the objects and their backgrounds. Unlike segmentation methods that perform pixel-wise comparisons to differentiate between foreground and background, object detectors omit this analysis, further aggravating the challenge. To solve this problem, we propose a camouflage-aware feature refinement (CAFR) strategy. Since camouflaged objects are not rare categories, CAFR fully utilizes a clear perception of the current object within the prior knowledge of large models to assist detectors in deeply understanding the distinctions between background and foreground. Specifically, in CAFR, we introduce the Adaptive Gradient Propagation (AGP) module that fine-tunes all feature extractor layers in large detection models to fully refine class-specific features from camouflaged contexts. We then design the Sparse Feature Refinement (SFR) module that optimizes the transformer-based feature extractor to focus primarily on capturing class-specific features in camouflaged scenarios. To facilitate the assessment of RCOD tasks, we manually annotate the labels required for detection on three existing segmentation COD datasets, creating a new benchmark for RCOD tasks. Extensive experiments on our proposed datasets demonstrate that the CAFR strategy significantly improves the model’s foreground and background recognition abilities and enhances RCOD task performance. Code and datasets are available at: https://***/zhimengXin/RCOD.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Secure Node Identification Through Covert Clock Feature Extraction

Secure Node Identification Through Covert Clock Feature Extr...

引用

IEEE Conference on computer communications Workshops, INFOCOM Wksps

作者： Xintao Huan Wen Chen Changfan Wu Jiamin Liu Yixuan Zou Shengkang Zhang School of Cyberspace Science and Technology Beijing Institute of Technology Beijing China School of Electronic Engineering and Computer Science Queen Mary University of London London U.K. Beijing Institute of Radio Metrology and Measurement Beijing China

ISBN: (数字)9798350384475

ISBN: (纸本)9798350384482

Node identification is the first line of defense for wireless networks' security that prevents illegal devices from accessing the network. Hardware features originating from innate hardware imperfections are considered promising identification fingerprints, among which, the hardware clock feature has been put under the spotlight due to its prevalence and ease of extraction. Current extractions of hardware clock features over wireless networks rely on the transmissions of time information, which, per se, face significant threats such as spoofing and replay attacks. In this paper, we propose a covert method to extract the hardware clock features which avoids the time information transmissions. We further propose a tailored secure node identification approach based on the proposed covert clock feature extraction. We implement and evaluate the proposed approach on a real Long Range (LoRa) testbed consisting of heterogeneous LoRa end nodes. Experimental results demonstrate that the proposed approach could effectively identify the heterogeneous LoRa end nodes with 95% accuracy.

关键词： Wireless networks LoRa Information processing Fingerprint recognition Feature extraction Hardware Data mining

来源：评论

学校读者我要写书评

暂无评论

A Novel Malware Detection Approach Based on Behavioral Semantic Analysis and LSTM Model 21

A Novel Malware Detection Approach Based on Behavioral Seman...

引用

21st IEEE International Conference on communication technology, ICCT 2021

作者： Han, Weijie Xue, Jingfeng Qian, Kechang Space Engineering University School of Space Information Beijing China School of Computer Science and Technology Beijing Institute of Technology Beijing China

ISBN: (纸本)9781665432061

Malware has been a major security threat to cyberspace. To identify the maliciousness of a program, researchers usually extract relevant semantic features from the program to analyze the operational intent of the program, and the most common object of analysis is the sequence of system calls of the program. System call sequences can effectively reflect the behavioral characteristics of a program, but are relatively low-level and poorly comprehensible, making it difficult to intuitively reflect the behavioral intent of a program. For this reason, this paper further analyzes the operation behavior types corresponding to the program system calls on the basis of the analysis of the program system call sequences, so as to realize the understanding of the program behavior semantics based on the analysis of the program behavior operation sequences. In this paper, 15 behavior types are defined to characterize the behavior semantics of the program, and the program feature vector is constructed by extracting the operation sequence of the program behavior types, which can simplify the complexity of the feature vector and understand the behavior operation intention of the program more accurately, and finally the detection accuracy of the classifier constructed based on the deep learning model LSTM can reach 96.14 %. © 2021 IEEE.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Improved YOLOv5 Algorithm for Intensive Pedestrian Detection 29th

Improved YOLOv5 Algorithm for Intensive Pedestrian Detection

引用

29th International Conference on Computational and Experimental engineering and sciences, ICCES 2023

作者： Zhao, Siqi Tian, Ye Hao, Ning Zhou, Jianbo Zhang, Xian College of Computer Science and Technology Harbin Engineering University Harbin150001 China Modeling and Emulation in E-Government National Engineering Laboratory Harbin150001 China Hangzhou Institute of Technology Xidian University Hangzhou311200 China Beijing Institute of Spacecraft Environment Engineering Haidian District Space City Beijing100048 China Hangzhou Hikvision Digital Technology Co. Ltd Hangzhou310051 China School of Physics Mathematics and Computing University of Western Australia PerthWA6009 Australia

ISBN: (纸本)9783031449468

Pedestrian detection is an important basic research topic in the field of target detection, which can provide effective information support for public places with large flow density such as shopping malls and scenic spots as well as intelligent security fields. To solve the problem of tracking target loss and low detection and recognition rate caused by pedestrian occlusion or portrait over-lap in pedestrian target detection in a crowded area or scene, an improved YOLOv5 algorithm integrating the attention mechanism was proposed. The relationship between feature channels and spatial information of the feature map was deeply mined by introducing an attention mechanism to further enhance feature extraction of pedestrian target visual area. To improve the feature fusion ability, Bidirectional Feature Pyramid Network (BiFPN) feature pyramid was used to enrich the cross-scale connection mode, preserve the shallow layer characteristics, and improve the detection accuracy. To improve the convergence ability of the model, EIoU was used to replace the original loss function of YOLOv5 to optimize the regression prediction of the anchor, which reduces the training difficulty of the network and improves the detection rate under occlusion. Compared with the general YOLOv5 algorithm, the improved algorithm proposed in this paper has higher accuracy and a lower missing rate in pedestrian detection in crowded areas or scenes, while the real-time performance of the original algorithm is still maintained. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Robust Dual-Modal Speech Keyword Spotting for XR Headsets

arXiv

引用

arXiv 2024年

作者： Cai, Zhuojiang Ma, Yuhan Lu, Feng State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and Engineering Beihang University Beijing China

While speech interaction finds widespread utility within the Extended Reality (XR) domain, conventional vocal speech keyword spotting systems continue to grapple with formidable challenges, including suboptimal performance in noisy environments, impracticality in situations requiring silence, and susceptibility to inadvertent activations when others speak nearby. These challenges, however, can potentially be surmounted through the cost-effective fusion of voice and lip movement information. Consequently, we propose a novel vocal-echoic dual-modal keyword spotting system designed for XR headsets. We devise two different modal fusion approches and conduct experiments to test the system’s performance across diverse scenarios. The results show that our dual-modal system not only consistently outperforms its single-modal counterparts, demonstrating higher precision in both typical and noisy environments, but also excels in accurately identifying silent utterances. Furthermore, we have successfully applied the system in real-time demonstrations, achieving promising results. The code is available at https://***/caizhuojiang/VE-KWS. Copyright © 2024, The Authors. All rights reserved.

关键词： Cost effectiveness

来源：评论

学校读者我要写书评

暂无评论

Adaptive Trajectory Planning in Autonomous Vehicles: A Hierarchical Reinforcement Learning Approach with Soft Actor-Critic

Adaptive Trajectory Planning in Autonomous Vehicles: A Hiera...

引用

International Symposium on Advanced Networks and Telecommunication Systems (ANTS)

作者： Amit Kumar Sharma Amit Choudhary Rajat Chaudhary Aditya Bhardwaj Anjum Mohd Aslam Electronics and Communication Sir Chhotu Ram Institute of Engineering and Technology Meerut Uttar Pradesh India Electrical Engineering and Computer Science KTH Royal Institute of Technology Stockholm Sweden School of Computer Science and Engineering Technology Bennett University Greater Noida India

ISBN: (数字)9798350391725

ISBN: (纸本)9798350391732

This study introduces a methodology enabling automated vehicles to perform lane changes effectively within complex road systems. It emphasizes a hierarchical driver behavior framework that integrates decision-making with trajectory planning to enhance safety. The approach utilizes reinforcement learning (RL) agents for automated vehicles and the MOBIL model for human-operated vehicles, aiming to optimize the lane change process. The paper introduces the Soft Actor-Critic (SAC), an off-policy actor-critic algorithm, to improve training stability and effectiveness in real-world robotics applications. Additionally, it offers a comprehensive review of existing planning and control algorithms for self-driving vehicles, offering a comprehensive survey of techniques and their strengths and limitations to aid in informed design choices.

关键词： Training Surveys Trajectory planning Roads Reinforcement learning Stability analysis Planning Telecommunications Safety Robots

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：