检索结果-内蒙古大学图书馆

2024 ieee International Performance, Computing, and Communications conference, IPCCC 2024

作者： Garcia-Luna-Aceves, J.J. Moghaddassian, Morteza Hsieh, Charles University of Toronto Department of Electrical and Computer Engineering Canada

ISBN: (纸本)9798350367942

Loop-free routing approaches based on distance information combine a local Ordering Condition (LOC) with a distributed reordering computation (DRC). The LOC allows routers to determine whether they can independently change next hops to destinations without creating a routing loop, and the DRC permits routers re-establish their ordering relative to a destination when the LOC is not satisfied. A new type of DRC called Spreading Computations is shown to be inherently more efficient than DRCs used in existing loop-free routing protocols. A new routing algorithm based on spreading computations is presented and compared with OSPF, DSDV, and RIPv2 using the ns-3 simulator. The results show that using spreading computations results in faster convergence speeds after link or router recoveries or failures. © 2024 ieee.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Text Watermarking-Based Federated Learning Backdoor Attack Detection Scheme 6

Text Watermarking-Based Federated Learning Backdoor Attack D...

引用

6th International conference on Next Generation Data-Driven networks, NGDN 2024

作者： Zuo, Yiming Yang, Zheng Chen, Dongdong Gu, Ke Jiang, Luyao Changsha University of Science & Technology School of Computer and Communication Engineering ChangSha China

ISBN: (纸本)9798350388374

With the widespread application of federated learning across various domains, backdoor attacks pose a serious threat to the security of models. In this paper, we proposes a text watermarking-based federated learning backdoor attack detection scheme. The core idea of the scheme is to use the aggregation server to embed and verify the text watermark in the process of global model aggregation, so as to achieve tracking and security protection of the global model. Secondly, in order to deal with the situation that malicious attackers may forge text watermarks, we also deployed a backdoor attack detection and defense based on local gradient symbols on the aggregation server. The scheme detects the gradient information uploaded by the local device, and only when the gradient symbols of most local models are consistent, can it be included in the aggregation process of the global model to ensure the security of the model. Finally, the experimental results verify the effectiveness of this method. © 2024 ieee.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

M-FANet: Multi-Feature Attention Convolutional Neural Network for Motor Imagery Decoding

引用

ieee TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING 2024年 32卷 401-411页

作者： Qin, Yiyang Yang, Banghua Ke, Sixiong Liu, Peng Rong, Fenqi Xia, Xinxing Shanghai Univ Sch Mech & Elect Engn & Automat Shanghai 200444 Peoples R China Shanghai Univ Res Ctr Brain Comp Engn Sch Mechatron Engn & Automat Sch Med Shanghai 200444 Peoples R China Minist Educ Engn Res Ctr Tradit Chinese Med Intelligent Rehabi Shanghai 201203 Peoples R China

Motor imagery (MI) decoding methods are pivotal in advancing rehabilitation and motor control research. Effective extraction of spectral-spatial-temporal features is crucial for MI decoding from limited and low signal-to-noise ratio electroencephalogram (EEG) signal samples based on brain-computer interface (BCI). In this paper, we propose a lightweight Multi-Feature Attention Neural Network (M-FANet) for feature extraction and selection of multi-feature data. M-FANet employs several unique attention modules to eliminate redundant information in the frequency domain, enhance local spatial feature extraction and calibrate feature maps. We introduce a training method called Regularized Dropout (R-Drop) to address training-inference inconsistency caused by dropout and improve the model's generalization capability. We conduct extensive experiments on the BCI Competition IV 2a (BCIC-IV-2a) dataset and the 2019 World robot conference contest-BCI Robot Contest MI (WBCIC-MI) dataset. M-FANet achieves superior performance compared to state-of-the-art MI decoding methods, with 79.28% 4-class classification accuracy (kappa: 0.7259) on the BCIC-IV-2a dataset and 77.86% 3-class classification accuracy (kappa: 0.6650) on the WBCIC-MI dataset. The application of multi-feature attention modules and R-Drop in our lightweight model significantly enhances its performance, validated through comprehensive ablation experiments and visualizations.

关键词： Brain-computer interface motor imagery convolutional neural networks multi-feature attention

来源：评论

学校读者我要写书评

暂无评论

A Hybrid CNN-Transformer Model for Efficient Hyperspectral Image Reconstruction 6

A Hybrid CNN-Transformer Model for Efficient Hyperspectral I...

引用

6th ieee International conference on Power, Intelligent Computing and Systems, ICPICS 2024

作者： Tian, Fenggang Ma, Long School of Computer Science and Engineering Shenyang Jianzhu University Shenyang China School of Mechanical Engineering Shenyang Jianzhu University Shenyang China

ISBN: (纸本)9798350374315

Recent research in hyperspectral image (HSI) reconstruction mainly emphasizes the development of intricate mappings through using convolutional neural networks (CNNs). Although CNNs excel at capturing local feature information, they struggle to effectively perceive global feature information, resulting in suboptimal reconstruction accuracy. An SRCTNet model was designed to solve the above problem, which integrates CNNs with Transformer modules to efficiently reconstruct HSIs from RGB images. The CNN block utilizes a multi-scale attention mechanism to reuse features and capture multi-level information, while the Transformer block utilizes self-attention mechanism to acquire global information. This fusion maximizes the retention of both local and global representations, ensuring high reconstruction accuracy while balancing computational complexity and expanding the receptive field. Evaluations on the challenging NTIRE2022 dataset show that SRCTNet excels in objective metrics such as mean relative absolute error (MRAE), root mean square error (RMSE), and peak signal-to-noise ratio (PSNR), outperforming six methods. Moreover, our model requires fewer parameters and computational resources, demonstrating its effectiveness and robustness in hyperspectral image reconstruction. © 2024 ieee.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Improving Motor Imagery-Based Brain-computer Interfaces with Chebyshev Graph Neural networks 11

Improving Motor Imagery-Based Brain-Computer Interfaces with...

引用

11th ieee International conference on computer Science and Network Technology, ICCSNT 2023

作者： Feng, Zhenzhou Zhang, Hang Yang, Jinhui Xiong, Xuedou Wang, Chaoyue Yang, Juan School of Physics and Electronics Shandong Normal University Shandong Jinan China

ISBN: (纸本)9798350311594

This study aims to enhance the accuracy and generalization of motor imagery-based brain-computer interface (MI-BCI) systems using a novel frequency-based graph convolutional neural network technique called Chebyshev Graph Neural networks (ChebGNNs). The study involves acquiring a substantial amount of motor imagery EEG signals and merging them with an open-source dataset to create a heterogeneous dataset. The study shows that ChebGNNs significantly improve the training and classification performance of the MI-BCI model. ChebGNNs outperform conventional frequency-based and time-domain techniques, achieving an average accuracy of 91.09%, 90.27%, and 89.24% on various datasets. The study demonstrates the promising potential of graph convolutional neural networks in improving the accuracy and generalization of MI-BCI systems. This enables a more effective capture of correlation information and local characteristics in signals, thereby enhancing the overall performance of the system. © 2023 ieee.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

LGCNet: Feature Enhancement and Consistency Learning Based on local and Global Coherence Network for Correspondence Selection

LGCNet: Feature Enhancement and Consistency Learning Based o...

引用

ieee International conference on Robotics and Automation (ICRA)

作者： Wu, Tzu-Han Chen, Kuan-Wen Natl Yang Ming Chiao Tung Univ Dept Comp Sci Hsinchu 300 Taiwan

ISBN: (纸本)9798350323658

Correspondence selection, a crucial step in many computer vision tasks, aims to distinguish between inliers and outliers from putative correspondences. The coherence of correspondences is often used for predicting inlier probability, but it is difficult for neural networks to extract coherence contexts based only on quadruple coordinates. To overcome this difficulty, we propose enhancing the preliminary features using local and global handcrafted coherent characteristics before model learning, which strengthens the discrimination of each correspondence and guides the model to prune obvious outliers. Furthermore, to fully utilize local information, neighbors are searched in coordinate space as well as feature space. These two kinds of neighbors provide complementary and plentiful contexts for inlier probability prediction. Finally, a novel neighbor representation and a fusion architecture are proposed to retain detailed features. Experiments demonstrate that our method achieves state-of-the-art performance on relative camera pose estimation and correspondence selection metrics on the outdoor YFCC100M [1] and the indoor SUN3D [2] datasets.

关键词： Statistics

来源：评论

学校读者我要写书评

暂无评论

IDTransformer: Transformer for Intrinsic Image Decomposition

IDTransformer: Transformer for Intrinsic Image Decomposition

引用

ieee/CVF International conference on computer Vision (ICCV)

作者： Das, Partha Gevers, Maxime Karaoglu, Sezer Gevers, Theo Univ Amsterdam Amsterdam Netherlands Concordia Univ Quebec City PQ Canada 3D Universum Amsterdam Netherlands

ISBN: (纸本)9798350307443

The aim of intrinsic image decomposition (IID) is to recover reflectance and the shading from a given image. As different combinations are possible, IID is an under constrained problem. Previous approaches try to constrain the search space using hand crafted priors. However, these priors are based on strong imaging assumptions and fall short when these do not hold. Deep learning based methods learn the problem end-to-end from the data. But these networks lack any explicit information about the image formation model. In this paper, an IID transformer approach (IDTransformer) is proposed by learning photometric invariant attention, derived from the image formation model, integrated in the transformer framework. The combination of invariant features in both a global and local setting allows the network to not only learn reflectance transitions, but also to group similar reflectance regions, irrespective of the spatial arrangement. Illumination and geometry invariant attention is exploited to generate the reflectance map, while illumination invariant and geometry variant attention is used to compute the shading map. Enabling physics-based explicit attention allows the network to be trained on a relatively small dataset. Ablation studies show that adding invariant attention improves the performance. Experiments on the Intrinsic In the Wild dataset shows competitive results with competing methods. The project page with the code is available at https://***/***/.

关键词： attention computer vision Intrinsics Image decomposition physics based learning superpixels transformers

来源：评论

学校读者我要写书评

暂无评论

Diffeomorphic Image Registration with Neural Velocity Field 23

Diffeomorphic Image Registration with Neural Velocity Field

引用

23rd ieee/CVF Winter conference on Applications of computer Vision (WACV)

作者： Han, Kun Sun, Shanlin Yan, Xiangyi You, Chenyu Tang, Hao Naushad, Junayed Ma, Haoyu Kong, Deying Xie, Xiaohui Univ Calif Irvine Irvine CA 92717 USA Yale Univ New Haven CT 06520 USA

ISBN: (纸本)9781665493468

Diffeomorphic image registration, offering smooth transformation and topology preservation, is required in many medical image analysis tasks. Traditional methods impose certain modeling constraints on the space of admissible transformations and use optimization to find the optimal transformation between two images. Specifying the right space of admissible transformations is challenging: the registration quality can be poor if the space is too restrictive, while the optimization can be hard to solve if the space is too general. Recent learning-based methods, utilizing deep neural networks to learn the transformation directly, achieve fast inference, but face challenges in accuracy due to the difficulties in capturing the small local deformations and generalization ability. Here we propose a new optimization-based method named DNVF (Diffeomorphic Image Registration with Neural Velocity Field) which utilizes deep neural network to model the space of admissible transformations. A multilayer perceptron (MLP) with sinusoidal activation function is used to represent the continuous velocity field and assigns a velocity vector to every point in space, providing the flexibility of modeling complex deformations as well as the convenience of optimization. Moreover, we propose a cascaded image registration framework (Cas-DNVF) by combining the benefits of both optimization and learning based methods, where a fully convolutional neural network (FCN) is trained to predict the initial deformation, followed by DNVF for further refinement. Experiments on two large-scale 3D MR brain scan datasets demonstrate that our proposed methods significantly outperform the state-of-the-art registration methods.

关键词： Deformable models Learning systems Deep learning Image registration Three-dimensional displays Neural networks Brain modeling

来源：评论

学校读者我要写书评

暂无评论

Boosting vision transformers for image retrieval 23

Boosting vision transformers for image retrieval

引用

23rd ieee/CVF Winter conference on Applications of computer Vision (WACV)

作者： Song, Chull Hwan Yoon, Jooyoung Choi, Shunghyun Avrithis, Yannis Dealicious Inc Seoul South Korea Inst Adv Res Artificial Intelligence IARAI Vienna Austria Athena RC Maroussi Greece

ISBN: (纸本)9781665493468

Vision transformers have achieved remarkable progress in vision tasks such as image classification and detection. However, in instance-level image retrieval, transformers have not yet shown good performance compared to convolutional networks. We propose a number of improvements that make transformers outperform the state of the art for the first time. (1) We show that a hybrid architecture is more effective than plain transformers, by a large margin. (2) We introduce two branches collecting global (classification token) and local (patch tokens) information, from which we form a global image representation. (3) In each branch, we collect multi-layer features from the transformer encoder, corresponding to skip connections across distant layers. (4) We enhance locality of interactions at the deeper layers of the encoder, which is the relative weakness of vision transformers. We train our model on all commonly used training sets and, for the first time, we make fair comparisons separately per training set. In all cases, we outperform previous models based on global representation. Public code is available at https://***/dealicious-inc/DToP.

关键词： Training Location awareness Image retrieval computer architecture Self-supervised learning Image representation Transformers

来源：评论

学校读者我要写书评

暂无评论

Pixel-level Monocular Depth Estimation Based on Global-local Feature Fusion 6

Pixel-level Monocular Depth Estimation Based on Global-local...

引用

6th International conference on Communications, Information System and computer Engineering, CISCE 2024

作者： Tian, Daoqiang Wang, Bo Xie, Liping College of Computer Science and Technology Taiyuan University of Science and Technology Taiyuan China Chinese Academy of Sciences Institute of Automation Beijing China

ISBN: (纸本)9798350353174

To address the issue of insufficient extraction of global features in CNN monocular depth estimation networks, leading to scale ambiguity and object edge ambiguity in predicted depth maps, this paper presents a pixel-level monocular depth estimation method based on global-local feature fusion. Multiple pre-trained Transformer blocks are employed as encoders to extract global features at multiple scales. Additionally, a global area enhancement module based on pyramid pool is designed to increase the receptive field of the network, providing effective global context prior for in-depth scene analysis and enhancing the network's ability to obtain feature information. Furthermore, a feature fusion module based on large kernel convolution attention is proposed to better utilize global and local features such as shape and texture, refining semantic scenes and improving the global contours of depth maps. Finally, an adaptive distribution depth interval and estimation module are designed according to the requirements of depth features in different regions of the scene. The depth information is estimated by the probability distribution of pixels and linear combination of discrete pixels, resulting in clearer predicted depth maps with detailed and smooth object outlines. © 2024 ieee.

关键词： Convolution

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：