检索结果-内蒙古大学图书馆

OAAFormer:Robust and Efficient Point Cloud Registration Through Overlapping-Aware Attention in Transformer

Journal of computer science & technology 2024年第4期39卷 755-770页

作者： Jun-Jie Gao Qiu-Jie Dong Rui-An Wang Shuang-Min Chen Shi-Qing Xin Chang-He Tu Wenping Wang School of Computer Science and Technology Shandong UniversityQingdao 266237China School of Information and Technology Qingdao University of Science and TechnologyQingdao 266061China College of Engineering Texas A&M UniversityTexasTX 77843U.S.A.

In the domain of point cloud registration,the coarse-to-fine feature matching paradigm has received significant attention due to its impressive *** paradigm involves a two-step process:first,the extraction of multilevel features,and subsequently,the propagation of correspondences from coarse to fine ***,this approach faces two notable ***,the use of the Dual Softmax operation may promote one-to-one correspondences between superpoints,inadvertently excluding valuable ***,it is crucial to closely examine the overlapping areas between point clouds,as only correspondences within these regions decisively determine the actual *** these issues,we propose OAAFormer to enhance correspondence *** the one hand,we introduce a soft matching mechanism to facilitate the propagation of potentially valuable correspondences from coarse to fine *** the other hand,we integrate an overlapping region detection module to minimize mismatches to the greatest extent ***,we introduce a region-wise attention module with linear complexity during the fine-level matching phase,designed to enhance the discriminative capabilities of the extracted *** on the challenging 3DLoMatch benchmark demonstrate that our approach leads to a substantial increase of about 7%in the inlier ratio,as well as an enhancement of 2%-4%in registration ***,to accelerate the prediction process,we replace the Conventional Random Sample Consensus(RANSAC)algorithm with the selection of a limited yet representative set of high-confidence correspondences,resulting in a 100 times speedup while still maintaining comparable registration performance.

关键词： point cloud registration coarse-to-fine overlapping region feature matching Transformer

来源：评论

学校读者我要写书评

暂无评论

Estimating the State of Health for Lithium-ion Batteries:A Particle Swarm Optimization-Assisted Deep Domain Adaptation Approach

引用

IEEE/CAA Journal of Automatica Sinica 2023年第7期10卷 1530-1543页

作者： Guijun Ma Zidong Wang Weibo Liu Jingzhong Fang Yong Zhang Han Ding Ye Yuan The School of Mechanical Science and Engineering Huazhong University of Science and TechnologyWuhan 430074China the College of Electrical Engineering and Automation Shandong University of Science and TechnologyQingdao 266590China the Department of Computer Science Brunel University LondonUxbridgeMiddlesexUB83PHUnited Kingdom the School of Information Science and Engineering Wuhan University of Science and TechnologyWuhan 430081China the School of Artificial Intelligence and Automation Huazhong University of Science and TechnologyWuhan 430074China

The state of health(SOH)is a critical factor in evaluating the performance of the lithium-ion batteries(LIBs).Due to various end-user behaviors,the LIBs exhibit different degradation modes,which makes it challenging to estimate the SOHs in a personalized *** this article,we present a novel particle swarm optimization-assisted deep domain adaptation(PSO-DDA)method to estimate the SOH of LIBs in a personalized manner,where a new domain adaptation strategy is put forward to reduce cross-domain distribution *** standard PSO algorithm is exploited to automatically adjust the chosen hyperparameters of developed DDA-based *** proposed PSODDA method is validated by extensive experiments on two LIB datasets with different battery chemistry materials,ambient temperatures and charge-discharge *** results indicate that the proposed PSO-DDA method surpasses the convolutional neural network-based method and the standard DDA-based *** Py Torch implementation of the proposed PSO-DDA method is available at https://***/mxt0607/PSO-DDA.

关键词： Deep transfer learning domain adaptation hyperparameter selection lithium-ion batteries(LIBs) particle swarm optimization state of health estimation(SOH)

来源：评论

学校读者我要写书评

暂无评论

Design of a Scoring System for National Fitness Volunteer Services under Deep Learning 4

Design of a Scoring System for National Fitness Volunteer Se...

引用

4th International Conference on Evolutionary Computing and Mobile Sustainable Networks, ICECMSN 2024

作者： Singh, Abhimeet Singla, Raghav Sodhi, Pulkit Pandian, M. Thurai School of Computer Science and Engineering Vellore Institute of Technology Vellore India

This work is developed to establish a comprehensive, scientific, and reasonable national fitness volunteer service scoring system under a long short-term memory (LSTM) recurrent neural network (RNN) algorithm. The LSTM RNN algorithm-based architecture of a three-layer information processing system for human motion recognition is proposed, which includes a data acquisition layer, a data calculation layer, and a data application layer. The LSTM RNN model is verified on the public dataset PAMAP2. Under the LSTM RNN-based information processing system for human motion recognition, the national fitness volunteer service scoring system is established by using the literature method, the survey method, the analytic hierarchy process (AHP) method, and the Delphi method. The candidate indicators are screened regarding the dimension of "service quality", the indicators are screened by the Delphi method, and the judgment matrix weights and consistency of each included indicator are analyzed. It is found that the information processing system for human motion recognition based on the LSTM RNN algorithm can effectively identify different motion states, such as sitting, lying down, running, walking, and cycling. A national fitness volunteer service scoring system based on the quality of communal sports facilities, services, fitness environment, affiliated sports facilities, and mass fitness service personnel is established under the LSTM RNN-based human motion recognition information processing system. The system includes one number of(#) first-level indicator, five number of(#) second-level indicators, and thirty two number of (#) third-level indicators. In summary, the LSTM RNN-based human motion recognition information processing system can correctly mine the types of fitness exercises, and the established national fitness volunteer service scoring system provides a scientific reference for building a perfect national fitness service system. © 2025 The Authors. Published by Els

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

An Investigation on Open-RAN Specifications:Use Cases,Security Threats,Requirements,Discussions

引用

computer Modeling in engineering & sciences 2024年第10期141卷 13-41页

作者： Heejae Park Tri-HaiNguyen Laihyuk Park Department of Computer Science and Engineering Seoul National University of Science and TechnologySeoul01811Republic of Korea Faculty of Information Technology School of TechnologyVan Lang UniversityHo Chi Minh City70000Vietnam

The emergence of various technologies such as terahertz communications,Reconfigurable Intelligent Surfaces(RIS),and AI-powered communication services will burden network operators with rising infrastructure ***,the Open Radio Access Network(O-RAN)has been introduced as a solution for growing financial and operational burdens in Beyond 5G(B5G)and 6G networks.O-RAN promotes openness and intelligence to overcome the limitations of traditional *** disaggregating conventional Base Band Units(BBUs)into O-RAN Distributed Units(O-DU)and O-RAN Centralized Units(O-CU),O-RAN offers greater flexibility for upgrades and network ***,this openness introduces new security challenges compared to traditional *** existing studies overlook these security requirements of the O-RAN *** gain deeper insights into the O-RAN system and security,this paper first provides an overview of the general O-RAN architecture and its diverse use cases relevant to B5G and 6G *** then delve into specifications of O-RAN security threats and requirements,aiming to mitigate security vulnerabilities *** providing a comprehensive understanding of O-RAN architecture,use cases,and security considerations,thisworkserves as a valuable resource for future research in O-RAN and its security.

关键词： O-RAN architecture use cases security issues security requirements security discussions

来源：评论

学校读者我要写书评

暂无评论

An Encoding-Based Privacy-Preserving Security Control Strategy Under Stealthy Attacks

IEEE Transactions on Industrial Cyber-Physical Systems

引用

IEEE Transactions on Industrial Cyber-Physical Systems 2025年 3卷 208-217页

作者： Liu, Jian Wang, Shuailong Zha, Lijuan Liu, Jinliang Tian, Engang Nanjing University of Finance and Economics School of Computer and Artificial Intelligence Nanjing210023 China Nanjing Forestry University School of Science Nanjing210037 China Nanjing University of Information Science and Technology School of Computer Science Nanjing210044 China University of Shanghai for Science and Technology School of Optical-Electrical and Computer Engineering Shanghai200093 China

This paper focuses on the privacy-preserving security control issue for cyber-physical systems (CPSs) via an encoding-decoding communication scheme (EDCS). To considerably contemplate the delayed data acquisition and analysis, the integral measurement model is established, which is governed by the function related to system states within a certain period. The EDCS designed has a certain resistant to the non-ideal communication network with limited bandwidth capacity and fading channels. Then the controller is constructed by utilizing the decoding data. Furthermore, in the controller-to-actuator channel, assuming the stealthy false data injection attacks occur randomly and the injected pseudo data is associated with transmission data that has been eavesdropped. By virtue of deriving sufficient conditions, the CPS with EDCS can achieve input-to-state stability as well as the corresponding controller and observer gain parameters are acquired. Eventually, the effectiveness of the designed EDCS and control algorithm is illustrated by a case study involving an autonomous underwater vehicle system. © 2023 IEEE.

关键词： Encoding (symbols)

来源：评论

学校读者我要写书评

暂无评论

Age estimation from facial images based on Gabor feature fusion and the CIASO-SA algorithm

引用

CAAI Transactions on Intelligence technology 2023年第2期8卷 518-531页

作者： Di Lu Dapeng Wang Kaiyu Zhang Xiangyuan Zeng School of Measurement and Communication Engineering Harbin University of Science and TechnologyHarbinChina Cheriton School of Computer Science University of WaterlooWaterlooOntarioCanada

Aiming at the problem of long time-consuming and low accuracy of existing age estimation approaches,a new age estimation method using Gabor feature fusion,and an improved atomic search algorithm for feature selection is ***,texture features of five scales and eight directions in the face region are extracted by Gabor wavelet *** statistical histogram is introduced to encode and fuse the directional index with the largest feature value on Gabor ***,a new hybrid feature selection algorithm chaotic improved atom search optimisation with simulated annealing(CIASO-SA)is presented,which is based on an improved atomic search algorithm and the simulated annealing ***,the CIASO-SA algorithm introduces a chaos mechanism during atomic initialisation,significantly improving the convergence speed and accuracy of the ***,a support vector machine(SVM)is used to get classification results of the age *** verify the performance of the proposed algorithm,face images with three resolutions in the Adience dataset are *** the Gabor real part fusion feature at 48�48 resolution,the average accuracy and 1-off accuracy of age classification exhibit a maximum of 60.4%and 85.9%,*** results prove the superiority of the proposed algorithm over the state-of-the-art methods,which is of great referential value for application to the mobile terminals.

关键词： age estimation atom search algorithm feature selection Gabor feature simulated annealing

来源：评论

学校读者我要写书评

暂无评论

Dynamic Low-Rank Adaptation Based Pruning Algorithm for Large Language Models 7

Dynamic Low-Rank Adaptation Based Pruning Algorithm for Larg...

引用

7th International Conference on Pattern Recognition and Artificial Intelligence, PRAI 2024

作者： Li, Linfeng Guo, Lei School of Computer Science and Technology Xinjiang University Urumqi China School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu China

ISBN: (纸本)9798350350890

Deep neural networks have been getting better and better performance while the parameter size has become larger and larger, and have now entered the era of large language models (LLM). However, large language models with huge number of parameters are difficult to be deployed on edge devices with limited computing power and memory resources. Network pruning techniques can reduce the storage and computation requirements of large language models. Although Dynamic Low-Rank Adaptation (DyLoRA) allows efficient fine-tuning of LLM, the deployment of LLM is still hindered by large model sizes and computational costs. In this paper, we propose a DyLoRA-based pruning algorithm that uses gradient updating of DyLoRA module to define the importance of parameters, and then prune the least important parameters based on the resulting importance scores. Through experiments on commonly used evaluation datasets, it is demonstrated that the proposed pruning algorithm is able to achieve 96.60% of the performance of the original model with 20% of the parameters pruned, which outperforms the previous pruning algorithms. © 2024 IEEE.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

DAUNet: Detail-Aware U-Shaped Network for 2D Human Pose Estimation

引用

computers, Materials & Continua 2024年第11期81卷 3325-3349页

作者： Xi Li Yuxin Li Zhenhua Xiao Zhenghua Huang Lianying Zou College of Information and Artificial Intelligence Nanchang Institute of Science and TechnologyNanchang330108China School of Electrical and Information Engineering Wuhan Institute of TechnologyWuhan430205China School of Computer Science and Technology Hubei Business CollegeWuhan430079China

Human pose estimation is a critical research area in the field of computer vision,playing a significant role in applications such as human-computer interaction,behavior analysis,and action *** this paper,we propose a U-shaped keypoint detection network(DAUNet)based on an improved ResNet subsampling structure and spatial grouping *** network addresses key challenges in traditional methods,such as information loss,large network redundancy,and insufficient sensitivity to low-resolution *** is composed of three main ***,we introduce an improved BottleNeck block that employs partial convolution and strip pooling to reduce computational load and mitigate feature ***,after upsampling,the network eliminates redundant features,improving the overall ***,a lightweight spatial grouping attention mechanism is applied to enhance low-resolution semantic features within the feature map,allowing for better restoration of the original image size and higher *** results demonstrate that DAUNet achieves superior accuracy compared to most existing keypoint detection models,with a mean PCKh@0.5 score of 91.6%on the MPII dataset and an AP of 76.1%on the COCO ***,real-world experiments further validate the robustness and generalizability of DAUNet for detecting human bodies in unknown environments,highlighting its potential for broader applications.

关键词： Human pose estimation keypoint detection U-shaped network architecture spatial grouping mechanism

来源：评论

学校读者我要写书评

暂无评论

Natural language requirements testability measurement based on requirement smells

引用

Neural Computing and Applications 2024年第21期36卷 13051-13085页

作者： Zakeri-Nasrabadi, Morteza Parsa, Saeed School of Computer Engineering Iran University of Science and Technology Hengam St. Resalat Square Tehran Tehran Iran

Requirements form the basis for defining software systems’ obligations and tasks. Testable requirements help prevent failures, reduce maintenance costs, and make it easier to perform acceptance tests. However, despite the importance of measuring and quantifying requirements testability, no automatic approach for measuring requirements testability has been proposed based on the requirements smells, which are at odds with the requirements testability. This paper presents a mathematical model to evaluate and rank the natural language requirements testability based on an extensive set of nine requirements smells, detected automatically, and acceptance test efforts determined by requirement length and its application domain. Most of the smells stem from uncountable adjectives, context-sensitive, and ambiguous words. A comprehensive dictionary is required to detect such words. We offer a neural word embedding technique to generate such a dictionary automatically. Using the dictionary, we could automatically detect Polysemy smell (domain-specific ambiguity) for the first time in 10 application domains. Our empirical study on nearly 1000 software requirements from six well-known industrial and academic projects demonstrates that the proposed smell detection approach outperforms Smella, a state-of-the-art tool, in detecting requirements smells. The Precision and Recall of smell detection are improved with an average of 0.03 and 0.33, respectively, compared to the state-of-the-art. The proposed requirement testability model measures the testability of 985 requirements with a mean absolute error of 0.12 and a mean squared error of 0.03, demonstrating the model’s potential for practical use. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Model Combining Vision Transformer and Generative Pre-trained Transformer for Image Captioning

A Hybrid Model Combining Vision Transformer and Generative P...

引用

2025 International Conference on Pervasive Computational Technologies, ICPCT 2025

作者： Gupta, Hina Vishwesh, Jatangi Y D M V Prasada Jagadeesh, M. Sucharitha, M. Pradyothan, Vitta Sai Bennett University School of Computer Science Engineering & Technology Noida India

ISBN: (纸本)9798331508685

Creating descriptive captions for images is now becoming a mission-critical application area in the intersection of natural language processing and computer vision. This work provides the hybrid model VisionGPT2, combining Vision Transformer for encoding images and GPT2 for text generation, which enables accurate, coherent, and context-aware image descriptions. The model is tested over a dataset by using BLEU score metrics and has achieved a high BLEU score of 0.4877, which ascertains the strength of cross-modal architectures for image captioning. In this paper, the design and training process of VisionGPT2 has been discussed along with the benefits and challenges faced while proceeding and the potential applications the model has gained in performance for future use. © 2025 IEEE.

关键词： computer vision

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：