检索结果-内蒙古大学图书馆

GM-DF: Generalized Multi-Scenario Deepfake Detection

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Lai, Yingxin Yu, Zitong Yang, Jing Li, Bin Kang, Xiangui Shen, Linlin The School of Computing and Information Technology Great Bay University Dongguan523000 China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen518060 China Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen University Shenzhen518060 China The Guangdong Key Laboratory of Information Security The School of Computer Science and Engineering Sun Yat-sen University Guangzhou510080 China Computer Vision Institute School of Computer Science & Software Engineering Shenzhen Institute of Artificial Intelligence and Robotics for Society Guangdong Key Laboratory of Intelligent Information Processing National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen518060 China

Existing face forgery detection usually follows the paradigm of training models in a single domain, which leads to limited generalization capacity when unseen scenarios and unknown attacks occur. In this paper, we elaborately investigate the generalization capacity of deepfake detection models when jointly trained on multiple face forgery detection datasets. We first find a rapid degradation of detection accuracy when models are directly trained on combined datasets due to the discrepancy across collection scenarios and generation methods. To address the above issue, a Generalized Multi-Scenario Deepfake Detection framework (GM-DF) is proposed to serve multiple real-world scenarios by a unified model. First, we propose a hybrid expert modeling approach for domain-specific real/forgery feature extraction. Besides, as for the commonality representation, we use CLIP to extract the common features for better aligning visual and textual features across domains. Meanwhile, we introduce a masked image reconstruction mechanism to force models to capture rich forged details. Finally, we supervise the models via a domain-aware meta-learning strategy to further enhance their generalization capacities. Specifically, we design a novel domain alignment loss to strongly align the distributions of the meta-test domains and meta-train domains. Thus, the updated models are able to represent both specific and common real/forgery features across multiple datasets. In consideration of the lack of study of multi-dataset training, we establish a new benchmark leveraging multi-source data to fairly evaluate the models' generalization capacity on unseen scenarios. Both qualitative and quantitative experiments on five datasets conducted on traditional protocols as well as the proposed benchmark demonstrate the effectiveness of our approach. The codes will be available on https://***/laiyingxin2/GM-DF. Copyright © 2024, The Authors. All rights reserved.

关键词： Image reconstruction

Efficient Region-Aware Neural Radiance Fields for High-Fidelity Talking Portrait Synthesis

学校读者我要写书评

暂无评论

Efficient Region-Aware Neural Radiance Fields for High-Fidel...

International Conference on computer Vision (ICCV)

作者： Jiahe Li Jiawei Zhang Xiao Bai Jun Zhou Lin Gu School of Computer Science and Engineering State Key Laboratory of Software Development Environment Jiangxi Research Institute Beihang University School of Information and Communication Technology Griffith University RIKEN AIP The University of Tokyo

This paper presents ER-NeRF, a novel conditional Neural Radiance Fields (NeRF) based architecture for talking portrait synthesis that can concurrently achieve fast convergence, real-time rendering, and state-of-the-art performance with small model size. Our idea is to explicitly exploit the unequal contribution of spatial regions to guide talking portrait modeling. Specifically, to improve the accuracy of dynamic head reconstruction, a compact and expressive NeRF-based Tri-Plane Hash Representation is introduced by pruning empty spatial regions with three planar hash encoders. For speech audio, we propose a Region Attention Module to generate region-aware condition feature via an attention mechanism. Different from existing methods that utilize an MLP-based encoder to learn the cross-modal relation implicitly, the attention mechanism builds an explicit connection between audio features and spatial regions to capture the priors of local motions. Moreover, a direct and fast Adaptive Pose Encoding is introduced to optimize the head-torso separation problem by mapping the complex transformation of the head pose into spatial coordinates. Extensive experiments demonstrate that our method renders better high-fidelity and audio-lips synchronized talking portrait videos, with realistic details and high efficiency compared to previous methods. Code is available at https://***/Fictionarry/ER-NeRF.

关键词：

A Dimensional Structure based Knowledge Distillation Method for Cross-Modal Learning

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Si, Lingyu Dong, Hongwei Qiang, Wenwen Yu, Junzhi Zhai, Wenlong Zheng, Changwen Xu, Fanjiang Sun, Fuchun Science & Technology on Integrated Information System Laboratory Institute of Software Chinese Academy of Sciences University of Chinese Academy of Sciences China College of Engineering Peking University China Department of Computer Science and Technology Tsinghua University China

Due to limitations in data quality, some essential visual tasks are difficult to perform independently. Introducing previously unavailable information to transfer informative dark knowledge has been a common way to solve such hard tasks. However, research on why transferred knowledge works has not been extensively explored. To address this issue, in this paper, we discover the correlation between feature discriminability and dimensional structure (DS) by analyzing and observing features extracted from simple and hard tasks. On this basis, we express DS using deep channel-wise correlation and intermediate spatial distribution, and propose a novel cross-modal knowledge distillation (CMKD) method for better supervised cross-modal learning (CML) performance. The proposed method enforces output features to be channel-wise independent and intermediate ones to be uniformly distributed, thereby learning semantically irrelevant features from the hard task to boost its accuracy. This is especially useful in specific applications where the performance gap between dual modalities is relatively large. Furthermore, we collect a real-world CML dataset to promote community development. The dataset contains more than 10, 000 paired optical and radar images and is continuously being updated. Experimental results on real-world and benchmark datasets validate the effectiveness of the proposed method. Copyright © 2023, The Authors. All rights reserved.

关键词： Distillation

DFPL: Decentralized Federated Prototype Learning Across Heterogeneous Data Distributions

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Zhang, Hongliang Xu, Fenghua Yu, Zhongyuan Hu, Chunqiang Pang, Shanchen Wang, Xiaofen Yu, Jiguo School of Computer Science and Technology Qilu University of Technology Shandong Jinan250353 China Cyber Security Institute University of Science and Technology of China Anhui Hefei230026 China School of Information Science and Engineering Lanzhou University Gansu Lanzhou730000 China School of Big Data and Software Engineering Chongqing University Chongqing400044 China Shandong Qingdao266580 China School of Computer Science and Engineering University of Electronic Science and Technology of China Sichuan Chengdu611731 China

Federated learning is a distributed machine learning paradigm that enables the collaborative training of multiple clients through centralized model aggregation. However, standard federated learning relies on a centralized server, making it vulnerable to server failures. While existing solutions utilize blockchain technology to implement Decentralized Federated Learning (DFL), the statistical heterogeneity of data distributions among clients severely degrades the DFL’s performance. Driven by this issue, this paper proposes a decentralized federated prototype learning framework, named DFPL, which significantly improves the performance of distributed machine learning across heterogeneous data distributions. Specifically, our framework introduces prototype learning into DFL to address statistical heterogeneity, which greatly reduces the number of parameters exchanged between clients. Additionally, blockchain is embedded into our framework, enabling the training and mining processes to be implemented at each client. From a theoretical perspective, we provide convergence guarantee of DFPL by combining resource allocation for training and mining. The experiments highlight the superiority of our DFPL framework in communication efficiency and test performance across three benchmark datasets with heterogeneous data distributions. © 2025, CC BY.

关键词： Federated learning

Technical Report: On the Convergence of Gossip Learning in the Presence of Node Inaccessibility

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Liu, Tian Cui, Yue Hu, Xueyang Xu, Yecheng Liu, Bo Zhejiang Laboratory Hangzhou China Department of Computer Science and Computer Information Systems Auburn University at Montgomery Montgomery United States Department of Computer Science and Software Engineering Auburn University Auburn United States Institute of Rural Development Zhejiang Academy of Agricultural Sciences Hangzhou China DBAPPSecurity Co. Ltd. Hangzhou China

Gossip learning (GL), as a decentralized alternative to federated learning (FL), is more suitable for resource-constrained wireless networks, such as Flying Ad-Hoc Networks (FANETs) that are formed by unmanned aerial vehicles (UAVs). GL can significantly enhance the efficiency and extend the battery life of UAV networks. Despite the advantages, the performance of GL is strongly affected by data distribution, communication speed, and network connectivity. However, how these factors influence the GL convergence is still unclear. Existing work studied the convergence of GL based on a virtual quantity for the sake of convenience, which failed to reflect the real state of the network when some nodes are inaccessible. In this paper, we formulate and investigate the impact of inaccessible nodes to GL under a dynamic network topology. We first decompose the weight divergence by whether the node is accessible or not. Then, we investigate the GL convergence under the dynamic of node accessibility and theoretically provide how the number of inaccessible nodes, data non-i.i.d.-ness, and duration of inaccessibility affect the convergence. Extensive experiments are carried out in practical settings to comprehensively verify the correctness of our theoretical findings. Copyright © 2024, The Authors. All rights reserved.

关键词： Antennas

Resource Optimization for LEO Constellation Networks: A Multi-satellite Cooperative Coverage Design

学校读者我要写书评

暂无评论

IEEE Transactions on Communications 2025年

作者： Ma, Xu Zhang, Haijun Song, Wei Lu, Yang Wu, Yuan Leung, Victor C.M University of Science and Technology Beijing Beijing Engineering and Technology Research Center for Convergence Networks and Ubiquitous Services Beijing100083 China China Electric Power Research Institute Co. Ltd Beijing102209 China University of Macau State Key Laboratory of Internet of Things for Smart City Macau China University of Macau Department of Computer and Information Science Macau China Shenzhen University Department of Computer Science and Software Engineering Shenzhen518061 China

With its low latency, high throughput, good deployment flexibility, and cost-effectiveness, the low earth orbit (LEO) constellation is regarded as a promising technology for seamless coverage. Unlike geosynchronous orbiting (GSO) satellites, the highly dynamic evolution of LEO constellation topology poses a significant challenge to the traditional scheme of allocating wireless resources. This paper proposes a multisatellite cooperative coverage resource allocation for the LEO constellation. A multi-objective optimization problem focusing on throughput and coverage time is proposed for the overall service quality of the LEO constellation. To solve the complex coupling between multi-domain resources, the optimization problem is decomposed into three key subproblems, which are the beam placement problem, the joint beam association with power allocation problem, and the beam hopping time slot allocation problem. For each subproblem, an adaptive algorithm is proposed that effectively exploits the payload of LEO satellites to ensure the coverage performance of the constellation network. Simulation results demonstrate that the proposed coverage scheme enables fast convergence, enhances network throughput utility, and guarantees the stability and fairness of the communication services. © 1972-2012 IEEE.

关键词： Multiobjective optimization

Blockchain-based Medical Image Data Trading Platform with Copyright and Privacy Protection

学校读者我要写书评

暂无评论

Blockchain-based Medical Image Data Trading Platform with Co...

International Conference on computer Supported Cooperative Work in Design

作者： Baowei Wang Wenjue Huang Bin Li Yi Yuan Fan Yang Zhengyu Hu Engineering Research Center of Digital Forensics Ministry of Education Nanjing China Waterford Institute Nanjing University of Information Science and Technology Nanjing China School of Computer and Software Nanjing University of Information Science and Technology Nanjing China Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology (CICAEET) Nanjing China

ISBN: (数字)9798350349184

ISBN: (纸本)9798350349191

In recent years, blockchain-based data trading platforms have gained great popularity in many fields because blockchain is decentralized, transparent, antitamper, and traceable. However, data trading for medical images still has problems such as privacy protection and copyright protection, so we propose a blockchain-based data trading scheme for medical images with copyright and privacy protection. Our scheme utilizes smart contracts in the blockchain to achieve automation and intelligence of data trading. We also combine digital watermarking and image encryption to further solve copyright, secure transmission, and privacy issues during medical image transactions. We conduct relevant experiments and verify the feasibility of our proposed scheme.

关键词： Data privacy Privacy Federated learning Smart contracts Watermarking Pricing Encryption

Typical curve with G1 constraints for curve completion

学校读者我要写书评

暂无评论

Visual Computing for Industry,Biomedicine,and Art 2021年第1期4卷 283-294页

作者： Chuan He Gang Zhao Aizeng Wang Fei Hou Zhanchuan Cai Shaolin Li School of Mechanical Engineering and Automation Beihang UniversityBeijing 100191China State Key Laboratory of Lunar and Planetary Sciences Macao University of Science and TechnologyMacao 999078China State Key Laboratory of Virtual Reality Technology and Systems Beihang UniversityBeijing 100191China State Key Laboratory of Computer Science Institute of SoftwareChinese Academy of SciencesBeijing 100190China State Key Laboratory of Computer Science Institute of SoftwareUniversity of Chinese Academy of SciencesBeijing 100049China Faculty of Information Technology Macao University of Science and TechnologyMacao 999078China

This paper presents a novel algorithm for planar G1 interpolation using typical curves with monotonic *** G1 interpolation problem is converted into a system of nonlinear equations and sufficient conditions are provided to check whether there is a *** proposed algorithm was applied to a curve completion *** main advantages of the proposed method are its simple construction,compatibility with NURBS,and monotonic curvature.

关键词： Typical curves Monotonic curvature G1 interpolation Curve completion Euler spiral

SEMDA: Secure and Efficient Multidimensional Data Aggregation in Smart Grid without a Trusted Third Party

学校读者我要写书评

暂无评论

Security and Communication Networks 2023年第1期2023卷

作者： Song, Zichao Zhong, Weidong Zhou, Tanping Chen, Dong Ding, Yujie Yang, Xiaoyuan College of Cryptography Engineering Engineering University of PAP Xi'an710086 China Key Laboratory of PAP for Cryptology and Information Security Xi'an710086 China TCA Laboratory State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing100080 China

Smart grids are a combination of traditional power system engineering as well as information and communications technology. Smart grid terminals provide convenient services to users by aggregating their data in real time. However, terminals can derive user privacy information from real-time data on smart devices. Therefore, security data aggregation has been widely studied in the field of smart grid. Most existing schemes are one-dimensional data aggregation or rely on a trusted third party. In reality, multidimensional data (such as a user's electricity consumption or user's main usage time, etc.) makes sense for terminals to flexibly adjust supply and demand strategies. In this paper, we propose an efficient and secure multidimensional data aggregation scheme that supports batch validation without a trusted third party. Firstly, we apply the Chinese remainder theorem to encode the user's multidimensional data and realize the independence of each dimension in terminal decryption. Secondly, we adopt a secure key negotiation protocol that does not require a trusted third party. Finally, based on paillier homomorphic encryption and bilinear pairing, we construct an encryption scheme that can reuse the key and blind factor and support batch verification. The analysis results show that our scheme is secure for users' privacy protection. Experimental results show that, compared with existing 1 dimensional aggregation schemes, our scheme has almost no growth in computational overhead for terminal decryption. © 2023 Zichao Song et al.

关键词： Smart power grids