检索结果-南通市图书馆

Local Region Frequency Guided Dynamic Inconsistency Network for Deepfake Video Detection

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

BIG DATA MINING AND ANALYTICS 2024年第3期7卷 889-904页

作者： Yue, Pengfei Chen, Beijing fu, zhangjie Nanjing Univ Informat Sci & Technol Engn Res Ctr Digital Forens Minist Educ Nanjing 210044 Peoples R China Nanjing Univ Informat Sci & Technol Sch Comp Sci Nanjing 210044 Peoples R China Nanjing Univ Informat Sci & Technol Jiangsu Collaborat Innovat Centerof Atmospher Envi Nanjing 210044 Peoples R China

In recent years, with the rapid development of deepfake technology, a large number of deepfake videos have emerged on the Internet, which poses a huge threat to national politics, social stability, and personal privacy. Although many existing deepfake detection methods exhibit excellent performance for known manipulations, their detection capabilities are not strong when faced with unknown manipulations. Therefore, in order to obtain better generalization ability, this paper analyzes global and local inter-frame dynamic inconsistencies from the perspective of spatial and frequency domains, and proposes a Local region Frequency Guided Dynamic Inconsistency Network (LFGDIN). The network includes two parts: Global SpatioTemporal Network (GSTN) and Local Region Frequency Guided Module (LRFGM). The GSTN is responsible for capturing the dynamic information of the entire face, while the LRFGM focuses on extracting the frequency dynamic information of the eyes and mouth. The LRFGM guides the GTSN to concentrate on dynamic inconsistency in some significant local regions through local region alignment, so as to improve the model's detection performance. Experiments on the three public datasets (FF++, DFDC, and Celeb-DF) show that compared with many recent advanced methods, the proposed method achieves better detection results when detecting deepfake videos of unknown manipulation types.

关键词： Deepfakes Privacy Frequency-domain analysis Mouth Lighting Stability analysis Spatiotemporal phenomena deepfake video detection dynamic inconsistency local region local region frequency

Privacy-Preserving Content-Aware Search Based on Two-Level Index

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2019年第5期59卷 473-491页

作者： zhangjie fu Lili Xia Yuling Liu Zuwei Tian School of Computer&Software Nanjing University of Information Science&TechnologyNanjing210044China College of Computer Science and Electronic Engineering Hunan UniversityChina School of Information Science and Engineering Hunan First Normal UniversityHunanChina

Nowadays,cloud computing is used more and more widely,more and more people prefer to using cloud server to store ***,how to encrypt the data efficiently is an important *** search efficiency of existed search schemes decreases as the index *** solving this problem,we build the two-level ***,for improving the semantic information,the central word expansion is *** purpose of privacy-preserving content-aware search by using the two-level index(CKESS)is that the first matching is performed by using the extended central words,then calculate the similarity between the trapdoor and the secondary index,finally return the results in *** experiments and analysis,it is proved that our proposed schemes can resist multiple threat models and the schemes are secure and efficient.

关键词： Semantic search two-level index expanded central-keyword

Privacy-Preserving Genetic Algorithm Outsourcing in Cloud Computing

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Cyber Security 2020年第1期2卷 49-61页

作者： Leqi Jiang zhangjie fu College of Computer and Software Nanjing University of Information Science&TechnologyNanjing210044China Jiangsu Engineering Center of Network Monitoring Nanjing University of Information Science and TechnologyNanjing210044China

Genetic Algorithm(GA)has been widely used to solve various optimization *** the solving process of GA requires large storage and computing resources,it is well motivated to outsource the solving process of GA to the cloud ***,the algorithm user would never want his data to be disclosed to cloud ***,it is necessary for the user to encrypt the data before transmitting them to the *** the user will encounter a new *** arithmetic operations we are familiar with cannot work directly in the ciphertext *** this paper,a privacy-preserving outsourced genetic algorithm is *** user’s data are protected by homomorphic encryption algorithm which can support the operations in the encrypted *** is elaborately adapted to search the optimal result over the encrypted *** security analysis and experiment results demonstrate the effectiveness of the proposed scheme.

关键词： Homomorphic encryption genetic algorithm outsourcing cloud computing

Multi-UAV Cooperative GPS Spoofing Based on YOLO Nano

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Cyber Security 2021年第2期3卷 69-78页

作者： Yongjie Ding zhangjie fu Engineering Research Center of Digital Forensics Ministry of EducationNanjing University of Information Science and TechnologyNanjing210044China College of Information Science and Technology College of Cyber SecurityJinan UniversityGuangzhou510632China

In recent years,with the rapid development of the drone industry,drones have been widely used in many fields such as aerial photography,plant protection,performance,and *** effectively control the unauthorized flight of drones,using GPS spoofing attacks to interfere with the flight of drones is a relatively simple and highly feasible attack ***,the current method uses ground equipment to carry out spoofing *** attack range is limited and the flexibility is not *** on the existing methods,this paper proposes a multi-UAV coordinated GPS spoofing scheme based on YOLO Nano,which can launch effective attacks against target drones with autonomous movement:First,a single-attack drone based on YOLO Nano is *** target tracking scheme achieves accurate tracking of the target direction on a single-attack drone;then,based on the single-UAV target tracking,a multi-attack drone coordinated target tracking scheme based on the weighted least squares method is proposed to realize the target drone Finally,a new calculation method for false GPS signals is proposed,which adaptively adjusts the flight trajectory of the attacking drone and the content of the false GPS signal according to the autonomous movement of the target drone.

关键词： UAV safety GPS spoofing multi-UAV target detection

High Visual Quality Image Steganography Based on Encoder-Decoder Model

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Cyber Security 2020年第3期2卷 115-121页

作者： Yan Wang zhangjie fu Xingming Sun School of Computer and Software Nanjing University of Information Science and TechnologyNanjingChina

Nowadays,with the popularization of network technology,more and more people are concerned about the problem of cyber ***,a technique dedicated to protecting peoples’private data,has become a hot topic in the research ***,there are still some problems in the current *** example,the visual quality of dense images generated by some steganographic algorithms is not good enough;the security of the steganographic algorithm is not high enough,which makes it easy to be attacked by *** this paper,we propose a novel high visual quality image steganographic neural network based on encoder-decoder model to solve these problems mentioned ***,we design a novel encoder module by applying the structure of U-Net++,which aims to achieve higher visual ***,the steganalyzer is heuristically added into the model in order to improve the ***,the network model is used to generate the stego images via adversarial *** results demonstrate that our proposed scheme can achieve better performance in terms of visual quality and security.

关键词： Steganaography visual quality cyber security

Dynamic Resource Scheduling in Emergency Environment

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Information Hiding and Privacy Protection 2019年第3期1卷 143-155页

作者： Yuankun Yan Yan Kong zhangjie fu School of Computer&Software Nanjing University of Information Science&TechnologyNanjingChina Jiangsu Engineering Center of Network Monitoring School of Computer and SoftwareNanjingChina

Nowadays,emergency accidents could happen at any *** accidents occur unpredictably and the accidents requirements are *** accidents happen in a dynamic environment and the resource should be cooperative to solve the *** methods are focusing on minimizing the casualties and property losses in a static ***,they are lack in considering the dynamic and unpredictable event *** this paper,we propose a representative environmental model in representation of emergency and dynamic resource allocation model,and an adaptive mathematical model based on Genetic Algorithm(GA)to generate an optimal set of solution *** experimental results show that the proposed algorithm can get a set of better candidate solutions.

关键词： Cooperative allocation dynamic resource scheduling adaptive genetic algorithm

An LSTM-Based Malware Detection Using Transfer Learning

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Cyber Security 2021年第1期3卷 11-28页

作者： zhangjie fu Yongjie Ding Musaazi Godfrey School of Computer and Software Nanjing University of Information Science and TechnologyNanjing210044China Guangxi Key Laboratory of Cryptography and Information Security Guilin541004China College of Information Science and Technology College of Cyber SecurityJinan UniversityGuangzhou510632China

Mobile malware occupies a considerable proportion of *** the update of mobile device operating systems and the development of software technology,more and more new malware keep *** emergence of new malware makes the identification accuracy of existing methods lower and *** is an urgent need for more effective malware detection *** this paper,we propose a new approach to mobile malware detection that is able to detect newly-emerged malware ***,we build and train the LSTM-based model on original benign and malware samples investigated by both static and dynamic analysis ***,we build a generative adversarial network to generate augmented examples,which can emulate the characteristics of newly-emerged *** last,we use the augmented examples to retrain the 4th and 5th layers of the LSTM network and the last fully connected layer so that it can discriminate against newly-emerged *** experiments show that our malware detection achieved a classification accuracy of 99.94%when tested on augmented samples and 86.5%with the samples of newly-emerged malware on real data.

关键词： Malware detection long short term memory networks generative adversarial networks transfer learning augmented examples

A Two-Stage Highly Robust Text Steganalysis Model

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Cyber Security 2020年第4期2卷 183-190页

作者： Enlu Li zhangjie fu Siyu Chen Junfu Chen School of Computer and Software Nanjing University of Information Science and TechnologyNanjing210044China Guangxi Key Laboratory of Cryptography and Information Security Guilin541004China College of Information Science and Technology College of Cyber SecurityJinan UniversityGuangzhou510632China

With the development of natural language processing,deep learning,and other technologies,text steganography is rapidly ***,adversarial attack methods have emerged that gives text steganography the ability to actively spoof *** terrorists use the text steganography method to spread terrorist messages,it will greatly disturb social *** methods,especially those for resisting adversarial attacks,need to be further *** this paper,we propose a two-stage highly robust model for text *** proposed method analyzes and extracts anomalous features at both intra-sentential and inter-sentential *** the first phase,every sentence is first transformed into word *** obtain a high dimensional sentence vector,we use Bi-LSTM to obtain feature information for all words in the sentence while retaining strong *** the second phase,we input multiple sentences vectors into the GNN,from which we extract inter-sentential anomaly features and make a judgment as to whether the text contains secret *** addition,to improve the robustness of the model,we add adversarial examples to the training set to improve the robustness and generalization of the steganalysis ***,our proposed method is more robust and more accurate in detection compared to existing methods.

关键词： Text steganalysis adversarial attack natural language processing deep learning

A Review of Object Detectors in Deep Learning

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal on Artificial Intelligence 2020年第2期2卷 59-77页

作者： Chen Song Xu Cheng Yongxiang Gu Beijing Chen zhangjie fu School of Computer and Software Nanjing University of Information Science and TechnologyNanjingChina

Object detection is one of the most fundamental,longstanding and significant problems in the field of computer vision,where detection involves object classification and *** with the traditional object detection algorithms,deep learning makes full use of its powerful feature learning capabilities showing better detection ***,the emergence of large datasets and tremendous improvement in computer computing power have also contributed to the vigorous development of this *** the paper,many aspects of generic object detection are introduced and summarized such as traditional object detection algorithms,datasets,evaluation metrics,detection frameworks based on deep learning and state-of-the-art detection results for object ***,we discuss several promising directions for future research.

关键词： Object detection deep learning computer vision