检索结果-内蒙古大学图书馆

A Global-Local Parallel Dual-Branch Deep Learning Model with Attention-Enhanced Feature Fusion for Brain Tumor MRI Classification

引用

computers, Materials & Continua 2025年第4期83卷 739-760页

作者： Zhiyong Li Xinlian Zhou School of Computer Science and Engineering Hunan University of Science and TechnologyXiangtan411100China

Brain tumor classification is crucial for personalized treatment *** deep learning-based Artificial Intelligence(AI)models can automatically analyze tumor images,fine details of small tumor regions may be overlooked during global feature ***,we propose a brain tumor Magnetic Resonance Imaging(MRI)classification model based on a global-local parallel dual-branch *** global branch employs ResNet50 with a Multi-Head Self-Attention(MHSA)to capture global contextual information from whole brain images,while the local branch utilizes VGG16 to extract fine-grained features from segmented brain tumor *** features from both branches are processed through designed attention-enhanced feature fusion module to filter and integrate important ***,to address sample imbalance in the dataset,we introduce a category attention block to improve the recognition of minority *** results indicate that our method achieved a classification accuracy of 98.04%and a micro-average Area Under the Curve(AUC)of 0.989 in the classification of three types of brain tumors,surpassing several existing pre-trained Convolutional Neural Network(CNN)***,feature interpretability analysis validated the effectiveness of the proposed *** suggests that the method holds significant potential for brain tumor image classification.

关键词： Deep learning attention mechanism feature fusion dual-branch structure brain tumor MRI classification

来源：评论

学校读者我要写书评

暂无评论

Gallium-induced structural modification over gel-derived CuZnGaAlOx catalysts for methanol steam reforming

引用

Ceramics International 2025年第14期51卷 18775-18785页

作者： Huang, Ziru Guo, Chenxu Guo, Wenming Zhang, Yi Huang, Shaoqin Qin, Hang Xie, Wen Gao, Pengzhao Xiao, Hanning College of Materials Science and Engineering Hunan University Changsha410082 China School of Materials Science and Engineering Hunan University of Science and Technology Xiangtan411201 China

Methanol steam reforming (MSR) is an attractive way to supply hydrogen applied for proton exchange membrane fuel cells. However, there is still a challenge in designing high-activity, low CO selectivity and durable copper-based catalysts. Herein, Ga-promoted CuZnAlOx catalysts are prepared by the phase separation-assisted sol-gel method with nanosheet structures, which provides a desirable space for reactants to contact the catalyst surface. It is demonstrated that the incorporation of Ga3+ can modify the defective chemistry characteristics and reducibility of ZnO, thus increasing the oxygen vacancy defects and Cu+ species for facilitating MSR reaction. Experimental results prove that CZA-7.5 Ga catalyst attains 100 % methanol conversion and 0.015 % CO selectivity at 260 °C, exhibiting higher MSR conversion and lower CO selectivity compared with the Ga-free catalyst. CZA just reached a methanol conversion of 96.7 % with 0.027 % CO selectivity at 260 °C. Additionally, CZA-7.5 Ga catalyst displayed the catalytic stability with methanol conversion rate of 85.41 % and CO selectivity ranging from 0 to 0.04 % after the 100 h reaction, while the methanol conversion of CZA decreased to 69.43 %. The process of reactant and intermediate evolution on the catalyst surface was briefly discussed using FTIR spectroscopy. In this work, a design strategy for Ga-promoted CuZnAlOx catalysts synthesized via the method is proposed, which provides insights into designing the copper catalysts and the selection of structural promoters. © 2025 Elsevier Ltd and Techna Group S.r.l.

关键词： Sol-gel process

来源：评论

学校读者我要写书评

暂无评论

Instruction selection optimization for VLIW architecture based on classification node merging

引用

Journal of Supercomputing 2025年第7期81卷 1-19页

作者： Liu, Fangjun Zhang, Huifu Hu, Yonghua Xie, Anxing Cao, Wei Mo, Shangfeng School of Computer Science and Engineering Hunan University of Science and Technology XiangTan411201 China Hunan Key Laboratory for Service Computing and Novel Software Technology XiangTan411201 China

Traditional instruction selection methods fail to fully exploit the very long instruction word (VLIW) architecture’s efficient scalar instructions. We propose an optimized instruction selection method based on classification node merging. This method abstracts scalar intermediate instructions into nodes and divides them into associated and similar ones according to operational semantics. Combined with the VLIW architecture’s instruction set features and guided by the instruction cost model, the nodes are merged and optimized to select efficient instructions. We design an instruction generation framework that swiftly converts the low level virtual machine intermediate representation into the instruction sequence of the target architecture. The experimental results show that, compared with the instruction sequence generated by M7002-cc, the instruction sequence generated with the solution proposed in this paper exhibits an average execution performance that is up to 54% higher on the FT-M7002 hardware platform. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Real time systems

来源：评论

学校读者我要写书评

暂无评论

Coarse-to-fine lightweight meta-embedding for ID-based recommendations

引用

science China(Information sciences) 2025年第4期68卷 82-97页

作者： Yang WANG Haipeng LIU Zeqian YI Biao QIAN Meng WANG School of Computer Science and Information Engineering Hefei University of Technology College of Information and Intelligence Hunan Agricultural University

State-of-the-art recommender systems are increasingly focused on optimizing implementation efficiency, such as enabling on-device recommendations under memory constraints. Current methods commonly use lightweight embeddings for users and items or employ compact embeddings to enhance reusability and reduce memory usage. However, these approaches consider only the coarse-grained aspects of embeddings, overlooking subtle semantic nuances. This limitation results in an adversarial degradation of meta-embedding performance, impeding the system's ability to capture intricate relationships between users and items, leading to suboptimal recommendations. To address this, we propose a novel approach to efficiently learn meta-embeddings with varying grained and apply fine-grained meta-embeddings to strengthen the representation of their coarse-grained counterparts. Specifically, we introduce a recommender system based on a graph neural network, where each user and item is represented as a node. These nodes are directly connected to coarse-grained virtual nodes and indirectly linked to fine-grained virtual nodes, facilitating learning of multi-grained semantics. Fine-grained semantics are captured through sparse meta-embeddings, which dynamically balance embedding uniqueness and memory constraints. To ensure their sparseness, we rely on initialization methods such as sparse principal component analysis combined with a soft thresholding activation function. Moreover, we propose a weight-bridging update strategy that aligns coarse-grained meta-embedding with several fine-grained meta-embeddings based on the underlying semantic properties of users and items. Comprehensive experiments demonstrate that our method outperforms existing baselines. The code of our proposal is available at https://***/htyjers/C2F-MetaEmbed.

关键词： lightweight meta-embedding coarse-to-fine learning ID-based recommendations

来源：评论

学校读者我要写书评

暂无评论

DSCR: A Dynamic Secure Clustering Routing Scheme for UANETs Based on Reputation Mechanism

引用

IEEE Internet of Things Journal 2025年第10期12卷 14109-14123页

作者： Gong, Yinyan Chen, Yuxiang Li, Kuanching Liang, Wei Li, Xiong Wang, Jin Xiang, Yang Hunan University of Science and Technology School of Computer Science and Engineering Xiangtan411201 China Sanya Research Institute Hunan University of Science and Technology Sanya572024 China School of Computer Science and Engineering Hunan Key Laboratory for Service Computing and Novel Software Technology Xiangtan411201 China Institute for Cyber Security School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu611731 China Swinburne University of Technology School of Software and Electrical Engineering HawthornVIC3122 Australia

In UAV ad hoc networks (UANETs), rapid movement of nodes leads to frequent changes in network topology, increasing the risk of packet loss and affecting data transmission. Furthermore, current research on drone clustering lacks security considerations, which reduces the reliability of data transmission. Due to this, improving the stability and reliability of network data transmission in dynamically changing UANETs remains a challenge. In this work, we propose the reputation mechanism dynamically secure cluster routing scheme (DSCR) for UANETs, a design for cluster head election in UANETs using the evaluated value and reputation value of drones and forms clusters through this reputation mechanism to avoid malicious nodes from interfering with the clusters to improve the security of UANETs. We apply a residual link survival time-based intra- and inter-cluster algorithm based on residual link survival time for data forwarding of nodes, optimizing the dynamic routing strategy and reducing the packet loss rate of nodes. In addition, reinforcement learning is used in UANETs to achieve cluster decisions based on the current network state and cluster dynamics, effectively improving the stability of clusters. Experiments were conducted on the proposed method to verify its efficiency and stability. Compared to the ICRA, RICR, and DCA algorithms, the stability of the cluster is improved by 11.92%, 28.12%, and 75.2%, respectively, and the packet loss rate reduced by 41.48%, 44.17%, and 55.74%, demonstrating that DSCR is a compelling dynamic routing solution applicable to UANETs. © 2014 IEEE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

E2E-LANet: Unleashing the Potential of Linear Attention for High-Resolution Medical Image Segmentation via Pretraining Distillation 24

E2E-LANet: Unleashing the Potential of Linear Attention for ...

引用

5th International Symposium on Artificial Intelligence for Medicine science, ISAIMS 2024

作者： Tang, Hui Yu, Gang Cao, Lu School of Computer Science and Engineering Hunan University of Science and Technology Hunan Xiangtan China

ISBN: (纸本)9798400717826

Medical image segmentation plays an important role in computer-aid diagnosis. In the past years, convolutional neural networks, especially the UNet-based architectures with symmetric U-shape encoder-decoder structure and skip connection, have been widely used in a variety of medical imaging segmentation tasks. However, such methods are difficult to capture global context. In contrast, Vision Transformers (ViTs) utilize the self-attention mechanism to obtain long-range dependencies in images. Despite these advantages, the huge computational costs of the self-attention mechanism reduce the inference speed of models, thereby limiting their real-world applications. Meanwhile, hierarchical pixel-level decoding of U-shape architectures increases the complexity of large-size segmentation decoding. To address these limitations, in this paper we propose a linear attention-based end-to-end model for high-resolution medical image segmentation, named E2E-LANet. It consists of two modules: an image encoder based on constructed Efficient Vision Transformer (EVT) that reduces the high quadratic complexity of self-attention to linear complexity and a Query-Supported Mask Decoder (QSMD) that employs independent mask query for each type of medical images in segmentation mask predictions. It focuses on the detailed segmentation decoding of patch-level features to further improve the efficiency of large-size mask predictions. Especially, we leverage the advantages of pretraining distillation to improve the transferability of our model in various segmentation tasks. Experimental results on the nucleus segmentation task and the polyp segmentation task display the important role of pretraining distillation in improving the performance of our framework. E2E-LANet demonstrates superior performance compared to CNNs, ViTs, and latest Mamba-based U-shape architectures. © 2024 Copyright held by the owner/author(s).

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Nanoflower-like K-birnessite cathode material for high-capacity aqueous Zn-ion battery

引用

Ionics 2025年 1-10页

作者： Xiao, Yuxuan Han, Changxin Yin, Ting Zhou, Wenjing Chou, Juan Zheng, Yuhong Zhang, Fengyue Cheng, Juanjuan Ou, Yun Liu, Longfei School of Materials Science and Engineering Hunan University of Science and Technology Xiangtan411201 China Hunan Provincial Key Lab of Advanced Materials for New Energy Storage and Conversion Hunan University of Science and Technology Xiangtan411201 China

Aqueous zinc ion batteries (AZIBs) have become a research hotspot due to their advantages of low cost, high safety, and environmental friendliness. K-Birnessite (KxMnO2) has been proved to be a candidate cathode material. However, the low conductivity and capacity degradation issues of δ-MnO2 limit its application. In the study, KxMnO2 (x = 0.27 and 0.31) with a nanoflower structure to improve the specific capacity has been designed, and capacity deterioration related to the evolution of Zn4SO4(OH)6·xH2O was analyzed. The capacity of K0.27MnO2 and K0.31MnO2 were 401.8 and 412.4 mAh g−1 at 0.1 A g−1 in the initial cycle due to the large specific surface area provided by the nanoflower structure. After 100 cycles, the specific capacity of the K0.27MnO2 electrode was 156.8 mAh g−1 at 1.0 A g−1, with a capacity retention rate close to 80%. During cycling, Zn4SO4(OH)6·xH2O was formed on the surface of the KxMnO2 cathode and transformed from a thin slice to a cracked block, leading to slow ion transport. This work provides a perspective for high-performance cathode design in AZIBs. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Nanoflowers

来源：评论

学校读者我要写书评

暂无评论

Privacy-Preserving Publishing Mechanism for Real-Time Vehicle Trajectory in Vehicle-to-Grid Networks 5

Privacy-Preserving Publishing Mechanism for Real-Time Vehicl...

引用

5th IEEE International Conference on Power, Electronics and computer Applications, ICPECA 2025

作者： Wang, Yulong Xu, Jianbo Gong, Bo School of Computer Science and Engineering Hunan University of Science and Technology Xiangtan China

ISBN: (纸本)9798331533694

The Vehicle-to-Grid (V2G) network is a smart grid technology generated under the background of the rapid development of new energy technology, which allows mobile energy storage vehicles (MESVs) to realize bidirectional energy exchange with the grid system. As distributed energy storage units within the V2G network, MESVs help meet the grid's load-balancing demands. To accurately schedule grid loads, V2G networks rely on MESVs' real-time trajectory data. After collecting this data, the central authority (CA) publishes it to the grid enterprise (GM) for load-balancing decisions. However, before publication, the CA must process the data to protect the location privacy of MESVs and prevent GM from extracting sensitive information. To address this privacy protection challenge during data publication, we propose a differential privacy-based mechanism. Our mechanism uses the unscented Kalman filter to predict and correct the published data at the sampling point and intelligently adjusts the sampling interval. Additionally, we introduce a privacy budget allocation method based on regional access frequency and travel time, enhancing privacy protection of trajectory in high-access areas and peak hours. Through security analysis and experimental validation, our mechanism effectively preserves the privacy of real-time vehicle trajectory data while ensuring data validity. © 2025 IEEE.

关键词： Vehicle-to-grid

来源：评论

学校读者我要写书评

暂无评论

Research on Network Security Game Based on System Dynamics 24

Research on Network Security Game Based on System Dynamics

引用

2nd International Conference on Artificial Intelligence, Systems and Network Security, AISNS 2024

作者： Zeng, Qiufen Pei, Tingrui School of Mechanical Engineering and Mechanics Xiangtan University Hunan Xiangtan China School of Information Science and Engineering Hunan Institute of Engineering Hunan Xiangtan China College of Information Science and Technology Jinan University Guangdong Zhuhai China School of Computer Science School of Cyberspace Science Xiangtan University Hunan Xiangtan China

ISBN: (纸本)9798400711237

Network security game was usually considered as attack-defense confrontation;however, we need to think out of the box because any network entity may get a chance to launch an attack. The problem of network security which was described as the game between regulatory authority and network entity needs to be analyzed from the perspective of macro management. The model of system dynamic which was combined with traditional game theory was established. Several strategies which were most commonly used were verified and evaluated through simulation model. In order to weaken the instability of network security game, selective punishment strategy was proposed and the simulation result showed that not only in astringency but also in network security maintenance the proposed strategy can achieve good results. © 2024 Copyright held by the owner/author(s).

关键词： System theory

来源：评论

学校读者我要写书评

暂无评论

A robust multi-VSGs coordinated control strategy for distributed secondary frequency regulation in islanded microgrids

引用

science China(Technological sciences) 2025年第3期68卷 331-348页

作者： Jing ZHANG Yong LI Jinjie LIN Sijia HU Yijia CAO Fang LIU Mingmin ZHANG Li HE College of Electrical and Information Engineering Hunan UniversityChangsha410082China Greater Bay Area Institute for Innovation Hunan UniversityGuangzhou511300China College of Information Science and Engineering Central South UniversityChangsha410083China College of Electronic Information and Electrical Engineering Changsha UniversityChangsha410022China College of Information and Electrical Engineering Hunan University of Science and TechnologyXiangtan411201China

Islanded microgrids(IMGs)offer a viable and efficient energy self-sustaining solution for distributed resources in remote *** without utility grid support,the frequency of IMG is susceptible to mismatches between demand and ***,IMGs encounter uncertain and nonlinear load disturbances together with system parameter perturbation,which further compromises frequency *** this aim,this paper proposes a robust multi-virtual synchronous generators(multi-VSGs)coordinated control strategy for distributed secondary frequency regulation(DSFR)in IMGs,which exhibits minimal model dependency and avoids reliance on global *** critical methods are developed:(1)a robust VSG control framework that incorporates the linear active disturbance rejection control(LADRC)technique,which enables the estimation and effective elimination of uncertain load disturbances and system's parameter perturbations;(2)a novel secondorder consensus algorithm-based control law for robust secondary frequency regulation,which is featured with proper power sharing among different participants,suppressed power oscillation caused by response disparities,and reduced reliance on complex communication *** on methods(1)and(2),a novel multi-VSGs coordinated control strategy is proposed,providing a robust solution for IMG's frequency restoration,and its dynamic characteristics are explored in *** correctness and effectiveness of the proposal are verified by both simulation and the hardware-in-the-loop(HIL)experiment results across typical scenarios.

关键词： islanded microgrid(IMG) distributed secondary frequency regulation(DSFR) robust multi-VSGs coordinated control linear active disturbance rejection control(LADRC)

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：